close

Вход

Забыли?

вход по аккаунту

?

How to select a new CAD system

код для вставки
How to select a new CAD system
P-O Friman, J Wikner
Preface
This thesis work was done at CERN in Geneva, Switzerland during the
period October 1999 to April 2000. The original request from CERN was to
develop benchmark tests for a number of CAD programs. The tests were to
be used as part of a CAD system selection process. A description of how to
select a new CAD system with detailed descriptions of benchmark tests has
been developed and some benchmark trials carried out. This work is the
final part of the requirements for a Master of Science in Mechanical
Engineering with specialisation in Machine Design at LuleГҐ University of
Technology, Sweden.
Acknowledgments
We would like to thank the following people for helping us in our work and
for making this thesis possible:
Supervisors at CERN:
Antti Onnela, EP-TA1
Thomas Pettersson, EST-ISS
Examiner at LTU:
Peter Jeppsson
Additional and appreciated for their help during the work and stay at
CERN:
Kristina Gunne, Swedish Office
Financial support:
Naturvetenskapliga ForskningsrГҐdet
Per-Olof Friman and Jakob Wikner
Luleå – October 2000
How to select a new CAD system
P-O Friman, J Wikner
Abstract
This MSc. thesis project concerned how to choose a new CAD system in a
large organisation. Several aspects have been studied in detail, specifically
specifications of benchmark tests. The complete selection process is
described in the results, from the project start to the final decision.
This work has been carried out at CERN and all examples used refer to the
present situation at CERN concerning their use of CAD system. CERN are
currently evaluating their use of CAD systems in order to develop future
strategies as far as CAD is concerned. The project was started as a result of
the development of CERN’s main 3D CAD system stopping.
CERN is a large organisation and carries out much research and design in
collaboration with other companies and organisations around the world. The
demands on the CAD systems used at CERN are significant. Some
examples of the CAD systems requirements are:
• Reliable import and export of designs between different CAD programs.
• That designers should be able to work in parallel on the same design.
• Several different kinds of users, including temporary or casual CAD
users should be able to use powerful 3D CAD.
• The ability to handle large assemblies with 10’s of thousands of parts.
The strategy in this work was to first investigate how the CAD system
should be used as far as working methodologies and interaction with other
CAx/PDM systems was concerned. After this, requirements were specified
and evaluation of CAD systems can be carried out.
The emphasis in this work has been to stress requirements to organise and
manage data, and support possibilities to work in parallel according to the
principles of concurrent engineering. Another important area that has been
discussed is how data from the old system can be migrated to the new
system.
A large part of this thesis work has been to develop the benchmark method
with which to test CAD programs against the user’s needs. The resulting
method's objective is to test and compare different CAD programs against
one common test specification.
The work describes the complete process from project start to the
implementation of a new CAD system.
How to select a new CAD system
P-O Friman, J Wikner
Sammanfattning
Det här arbetet beskriver ämnet hur man väljer nytt CAD system i en
organisation. Flera aspekter Г¤r studerade i detalj och speciellt stor vikt har
lagts pГҐ test-specifikationer. I Resultatet finns sedan hela urvalsprocessen
beskriven, frГҐn projektstart till beslut.
Det här arbetet är utfört vid CERN och alla använda exempel bygger på
nuvarande situation. Vid CERN finns det nu ett behov av att byta CAD
system eftersom utvecklingen av CERN's primära 3D CAD system är
avbruten. DГҐ CERN Г¤r en stor organisation som har mycket forsknings- och
utvecklings samarbete med andra institutioner och företag runt om i världen,
ställs det stora krav på CAD systemet. Några exempel på detta är:
• Import och export av data ska kunna ske mellan olika CAD program.
• Flera designers ska kunna jobba på samma design parallellt.
• Många olika typer av användare, även tillfälliga användare ska kunna
använda kraftfull 3D CAD.
• Stora sammanställningar med 10 000 -tals delar ska kunna hanteras.
Strategin är att först undersöka vad man ska använda CAD-systemet till med
avseende pГҐ vilka arbetsmetodologier och CAx/PDM system man ska
använda. Sedan kan man utvärdera de olika programmen efter de krav och
önskemål man har ställt upp.
Olika aspekter vid val av CAD system är behandlade där stor tonvikt är lagd
på möjligheter vid organisering av data samt möjlighet till parallellt arbete
enligt principerna för “concurrent engineering”. Ett annat viktigt område
som ocksГҐ behandlats Г¤r hur data frГҐn andra och nuvarande system kan
överföras till det nya systemet.
En stor del av arbetet är utvecklingen av en metod för test av CAD program.
Detta för att se vilket program som bäst klarar av organisationens krav. Det
har resulterat i en metod för hur man kan testa och jämföra olika CAD
program med en och samma test specifikation.
Arbetet beskriver hur en generell CAD-urvalsprocess kan utföras från
projekt start till införandet av ett nytt CAD system.
How to select a new CAD system
P-O Friman, J Wikner
Glossary
This glossary explains many of the technical words, phrases and
abbreviations used in this paper. Since many design software vendors have
their own vocabulary, the expressions and abbreviations presented may
differ slightly from one vendor to another.
2D
3D
AEC
API
Assembly
Assembly-drawing
Assembly-modelling
Benchmark
Boundary
Representation
CAD
CADIM/EDB
CAE
CAM
CAx
CDD
CERN
Check in
Check out
CNC
Copy
DDE
Two-dimensional.
Three-dimensional.
Architecture, Engineering and Construction.
Application Programming Interface.
Gathering of parts and subassemblies to make one
unique assembled product.
A drawing that can be created on the CAD system to
represent the complete product or a major
subdivision.
The process by which individual solid models are
brought together to form an assembly model.
A standardised problem or test that serves as a basis
for evaluation and comparison of different products.
A geometric database method that defines and stores
a solid as a set of vertices, edges, and faces (points,
lines, curves and surfaces) which completely enclose
its volume.
Computer-Aided Design (in the context of this
paper: Mechanical CAD).
CERN's current Engineering Data Management
System. Developed by Eigner+Partner.
Computer
Aided
Engineering;
calculation,
simulation etc.
Computer-Aided Manufacturing.
Computer Aided “x”; A term covering CAD, CAM,
CAE tools etc.
CERN Drawing Directory; a drawing archive and
management system, based on Oracle and using
HPGL files.
European Laboratory for Particle Physics.
When an item is passed to the library. This is known
as “check in.” It is also possible to store different
versions of a part as modifications are made.
An item is “checked out” so that changes can be
made. Only one user at a time can check out a part to
make changes; although parts can usually be used
for reference purposes, see “Reference”.
Computer Numerical Control.
Create an unassociated local copy of an item.
Dynamic Data Exchange.
How to select a new CAD system
DMU
EDMS
EPS
ESA
FEA
FTP
GIF
GIS
GUI
HPGL
Item
Interface
JPG/JPEG
LAN
Library
Object
OLE
Open Architecture
Part
PDF
PDM
PI
Project
PS
P-O Friman, J Wikner
Digital Mock-up, a digital structural model built
accurately to scale chiefly for study, testing, or
display.
Engineering Data Management System (also called
PDM, Product Data Management).
Encapsulated PostScript.
European Space Agency.
Finite Element Analysis.
File Transfer Protocol.
Graphical Interchange Format, a file format for
bitmapped images.
Geographical Information System.
Graphical User Interface.
Hewlett Packard Graphics Language, a language
used for vector graphics, traditionally created as plot
files for pen plotters.
A part, assembly, drawing or FE model.
The point at which two systems interact with each
other. The interface can be a piece of hardware, a
common area of computer storage or some common
instruction shared by two or more programs. An
interface can also refer to hardware, including
device driver software that connects peripheral
devices to computer platforms and to networks.
Interfacing also covers network and operating
protocols, codes and operating standards.
Joint Photographic Experts Group, a standard for
compressing gray-scale or color still images.
Local Area Network.
Library is the term used in CAD software to describe
where items are stored in the database. Each project
can have one or more libraries.
An item that shares certain characteristics with other
items.
Object Linking and Embedding.
The incorporation of standard interfacing features in
both the hardware and software of a computer
system, allowing communication with other
components that incorporate the same standard
interfacing features.
An object that consist of lines, surfaces or solids.
Portable Document Format.
Product Data Management.
Price Inquiry.
When working, users belong to one or more specific
projects where they can work simultaneously with
the same project-libraries.
PostScript.
How to select a new CAD system
Reference
Revision
SAT
STEP
Subassembly
SVG
Team
TIFF
UR
URD
RI
TS
XML
Version
VRML
WWW
P-O Friman, J Wikner
An object from the library to use for reference only.
It is not possible to make changes to the reference
object, however, the reference can be updated if
changes are made to the original object.
Revisions are used to keep track of released versions
of a product or component.
Internal data format of the ACIS geometry modeller
which is used by several CAD/CAM systems.
ISO Standard for the Exchange of Product Data.
A sub-level assembly that is used in an Assembly
structure.
Scalable Vector Graphics;an XML standard.
A team refers to a group of users working on the
same project.
Tag Image File Format.
User Requirements.
User Requirements Document.
Request for Information.
Technical Specification.
eXtensible Markup Language.
When an item has been checked out, modified and is
checked in again, it can be stored in the library as a
new version.
Virtual Reality Modelling Language.
World Wide Web.
How to select a new CAD system
P-O Friman, J Wikner
Table of Content
Page
1 Introduction
1.1 CERN
1.2 CAD systems at CERN
1
1
2
2 Task description
3
3 Methodology
3.1 Project plan
3.2 Information gathering and preparatory studies
3.2.1 Request for Information
3.2.2 User Requirement Document
3.3 Going from URD to Benchmarks
4
4
5
6
6
6
4 Aspects to consider in the CAD selection process
4.1 Design aspects
4.1.1 Handling large assemblies
4.1.2 Concurrent engineering
4.1.3 CAD system integration with PDM tools
4.1.4 Transfer of data between CAx programs
4.2 Economical considerations
4.2.1 Negotiation strategies
4.2.2 Use of licenses
4.3 Hardware implementation
8
8
8
10
12
16
18
19
19
20
5 The process of selecting a new CAD system
5.1 Committee responsible for selecting CAD system
5.2 The CAD evaluation team
5.3 Planning
5.4 Analysis of the organisation
5.5 User Requirement Document
5.6 Technical specification
5.7 Request for information
5.8 Benchmark tests
5.8.1 User requirements to test
5.8.2 Standard table for tests
5.8.3 Hardware specification
5.8.4 Operating System requirements
5.8.5 Test objects used in test
5.8.6 Data exchange between CAx systems
5.8.7 Analysing the benchmark test results
5.9 Price inquiry
5.10 Final analyse of Technology and price aspects
5.11 Gather team to negotiate with vendors
5.12 Implementing the new CAD system
22
22
22
23
25
26
27
28
32
32
33
34
34
35
35
39
44
44
44
45
6 Conclusions
46
References
48
How to select a new CAD system
Appendices
Appendix 1
Appendix 2
Appendix 3
Appendix 4
P-O Friman, J Wikner
Number of pages
An example of a Benchmark
category: PDM integration
table,
An example of a Benchmark
category: 2D drawings
table,
An example of a Benchmark
category: Manage assemblies
table,
11
4
Example tables of drawing and geometry
libraries to be used in benchmarks
5
2
How to select a new CAD system
Table of figures
P-O Friman, J Wikner
Page
Figure 1, CERN's accelerators ....................................................................... 1
Figure 2, Flowchart of the main activities in CERN’s CAD 2000
project. ................................................................................................... 5
Figure 3, Enterprise wide PDM updated from shared network disc............ 12
Figure 4, CAD with link to enterprise-wide PDM....................................... 13
Figure 5, CAD with local PDM system linked to the main PDM
system .................................................................................................. 14
Figure 6, Main PDM system connected to a local PDM system using
a high-end CAD system (CAD I) and a mid-range CAD system
(CAD II)............................................................................................... 15
Figure 7, Using a PDM system tightly integrated with the local PDM
system in order to make the connection to existing enterprise
wide PDM system. ............................................................................... 16
Figure 8, Transfer of data via a third CAx program. ................................... 18
Figure 9 ,CAD evaluation flowchart over the main project blocks. ............ 24
Figure 10, CAD system licences.................................................................. 31
Figure 11, Characteristics of CAD companies............................................. 31
Figure 12, Gathering of CAx transfer results............................................... 39
Figure 13, Result tree for CAD system A.................................................... 40
Figure 14, result tree for CAD system B. .................................................... 41
Figure 15, Total results of fulfilled test steps for CAD systems A and
B. .......................................................................................................... 42
Figure 16, Category results of fulfilled test steps for systems A and
B. .......................................................................................................... 42
Figure 17, Group results of fulfilled test steps for systems A and B. .......... 43
Table 1, An example of a User Requirement............................................... 27
Table 2, An example of a User Requirement table with comments on
tests............................................................................................ 27
Table 3, Qualification matrix ....................................................................... 30
Table 4, An example of a test table.............................................................. 33
Table 5, test results from CAx transfer. ....................................................... 37
Table 6, Criterion weight factor (Wc) analyse. ............................................ 38
How to select a new CAD system
P-O Friman, J Wikner
0
How to select a new CAD system
P-O Friman, J Wikner
1 Introduction
CERN is the European Organisation for Nuclear Research, the world's
largest particle physics centre. CERN is currently in a situation where it’s
main 3D CAD software, which is no longer being developed commercially,
is to be replaced. A new overall CAD system solution is thus needed at
CERN.
1.1 CERN
CERN explores what matter is made of, and the forces that hold it together
[9].
In 1951, a provisional body was created called the "Conseil EuropГ©en pour
la Recherche NuclГ©aire" (CERN). In 1953 the Council decided to build a
central laboratory near Geneva. At that time, physics research concentrated
on understanding the insides of the atom. The work at the laboratory, in
line with particle physics research, has moved into higher and higher
energy densities. The Laboratory provides state-of-the-art scientific
facilities for researchers to use. There are now 20 member states and
around 6500 scientists, from 500 universities and over 80 nationalities, that
use the facilities at CERN for their research.
By accelerating particles to very high energies and smashing them into
targets or into each other, physicists can understand the nature of the
particles and their interactions. CERN's accelerator complex is the most
versatile in the world. It includes particle accelerators and colliders that
can handle beams of electrons, positrons, protons, antiprotons, and "heavy
ions" (the nuclei of atoms, such as oxygen, sulphur, and lead). Each type
of particle is produced in a different way, but passes then through a similar
succession of acceleration stages, moving from one machine to another.
The first steps are usually provided by linear accelerators, followed by
larger circular machines. CERN has 10 accelerators altogether, the largest
are the Large Electron Positron collider (LEP) and the Super Proton
Synchrotron (SPS), see Figure 1, CERN's accelerators. The circumference
of the LEP tunnel is 27 km and is situated about 100 meters under ground.
Figure 1, CERN's accelerators
1
How to select a new CAD system
P-O Friman, J Wikner
The LEP accelerator began operation in the summer of 1989 and will be
disassembled at the end of year 2000, to be replaced by the new Large
Hadron Collider (LHC).
The LHC will be installed in the existing LEP tunnel and will allow
proton-proton collisions at energies 10 times greater than any previous
machine. This will allow even smaller particles to be identified and
studied. The construction of the LHC was approved in December 1994 and
is planned to start operation by the year 2005.
1.2 CAD systems at CERN
Since the early 1980’s CAD systems have been used at CERN for
designing the mechanical structures used in the accelerators and
experiments.
Today the main 3D CAD system used for mechanical engineering at
CERN is Euclid 3 with a total of 80 licenses. There are also some 150
floating licenses of AutoCAD and Mechanical Desktop and 8
Pro/Engineer licenses.
Development of Euclid ceased with the last version, Euclid 3.2. A new
CAD system must therefore be selected to gradually take over the work
currently done using Euclid; and the other CAD systems as well. To
support this transition, a working group was created in 1999, called the
CAD 2000 Task Force. The group currently has the task of proposing a
CAD system to use in the future at CERN.
The long term aim is not only to replace Euclid but also to find a global
CAD solution for use at CERN. It must therefore address the work
currently done by designers using AutoCAD and Pro/Engineer. The new
CAD system must support a transition and transfer of data from the
existing systems. It is also important to look beyond basic CAD
functionality since CERN have a multi system environment. In addition to
CAD, numerous other engineering applications for analysis and design
support are used at CERN. The new CAD system solution is expected to
include modules covering the functionality of many of these different
applications.
2
How to select a new CAD system
P-O Friman, J Wikner
2 Task description
The CAD selection process at CERN has been defined by the CAD 2000
Task Force and can be divided into a number of sub-projects which will
lead to the implementation of the future CAD system. These projects have
resulted in a number of working documents:
• The User Requirements Document (URD) was developed during 1999
and 2000. This covers the basic guideline for the final CAD system.
The URD is a living document, which will be continuously developed
as the CAD 2000 project progress.
• The Request for Information (RI) is the document used when first
contacting system vendors to get information about potential CAD
programs.
• Benchmark specifications. These are based on the URD and
knowledge of CERN’s present working practices and defines methods
for testing potential CAD systems. These specifications are currently
being developed.
• Price Inquiry (PI). When it is clear which CAD systems and
implementations that are suitable for CERN, a PI shall be made.
The sub-projects also cover the actual process of testing the candidate
CAD systems, selection of the final system and implementation.
When this work was started, the main steps regarding how to perform the
selection process had already been defined by the CAD 2000 Task Force.
The requirements from CERN’s CAD users were gathered in a draft of the
User Requirement Document (URD) and a Request for Information (RI)
document had been prepared as the basis for the first contact with vendors.
This final year thesis concerns the process of CAD selection within
CERN’s CAD 2000 project.
The aim of this work was to develop a general description of how to
perform a CAD evaluation. The work also covered the evaluation
procedure from project start to the implementation of the new CAD
software. Of particular interest was the development of a benchmark
specification. The results of the work were, as far as possible, to be
applicable to the CERN CAD selection process whilst at the same time
satisfying the academic requirements set by LuleГҐ University of
Technology for a final year project.
3
How to select a new CAD system
P-O Friman, J Wikner
3 Methodology
An outline project plan was created at an early stage. The work began
with the gathering of information from which ideas of how to test CAD
systems developed. Discussions and trials lead to the final test
specifications for testing CAD systems.
3.1 Project plan
All projects start with some kind of planning activity. The project plan,
however, must be considered as a dynamic and not a static framework.
This was certainly the case in this project and the project plan was
continuously refined.
The first task was to create a rough plan for the project. This plan was
created during the first weeks of the project when pre-studies and
organisational analysis dominated the work. After studying CERN’s
current purchasing procedures for new products, a more detailed project
plan was made. As the project evolved, the plan was continuously revised
to take account of changing situations and new information. An overview
of the process of selecting a CAD system at CERN can be seen in the flow
chart shown in Figure 2, Flowchart of the main activities in CERN’s CAD
2000 .
4
How to select a new CAD system
P-O Friman, J Wikner
CAD
Evaluation
RI
Analyse
replies on
RI
List of programs
for tendering &
further studies
Benchmark
tests
Preparation
of Bench
Marks
Evaluation
of CAD
user needs
Benchmark
Specification
URD
Comm
ercial
Docs
Commercial
and
Technical
Specifications
Price
Inquiry
Analyse
Replies on
PI and BM results
Recommendation
of future CAD
Program(s) for CERN
Decision by
CERN Direction
NEXT CAD
PROGRAM
Figure 2, Flowchart of the main activities in CERN’s CAD 2000 project.
3.2 Information gathering and preparatory studies
In order to maximise the benefits of moving to new CAD systems in an
organisation, it is important to understand the system functionality and
working methodologies associated with the leading systems in the CAD
market.
Discussions with people at CERN gave information about their work and
experiences with the present CAD systems. In parallel, CAD publications
such as CAD Report, Computer Graphics World and other resources on
5
How to select a new CAD system
P-O Friman, J Wikner
the Internet were studied. Other main sources used when finding
background information was the URD and the vendor responses to the RI.
To find out about CERN’s requirements, aspects of the existing
organisation were studied. For example, the typical size of the projects and
the required co-operation between teams. There was also additional input
from the users and “in house” support. A large amount of information was
also obtained from the User Requirements Document.
3.2.1 Request for Information
The Request for Information (RI) was used by the CAD 2000 team when
contacting CAD vendors and to allow consistent information about the
companies and their products to be obtained. Following, the retrieved data
was subsequently used by the CAD 2000 Task Force to identify CAD
vendors and CAD systems worth further investigation. The vendor
responses were analysed and summarised in an internal CERN document.
3.2.2 User Requirement Document
Key persons in the CAD 2000 project had already created the first versions
of the User Requirement Document (URD) when this thesis work began.
English and French versions of the URD were sent out to a large group of
CERN CAD users (including experts) to get more feedback on the
requirements in the document. The feedback was then used to develop a
revised URD.
Gathering opinions from a large group of people was not without it’s
problems. The main advantage of consulting a large group of people is that
many requirements and opinions can be gathered; in this case based on
experience from present CAD systems. However, problem can occur when
interpreting feedback if individuals have differing opinions. This can make
some requirements appear contradictory. Experience from the URD
development process helped in writing the description of a CAD system
selection process.
3.3 Going from URD to Benchmarks
The URD expresses the needs of the users. All the requirements in the
URD were commented with additional explanations and suggestions for
test instructions. This formed the basis for defining how the tests should be
done in the benchmark.
All User Requirements (UR’s) were analysed in order to separate
them into those to test and those not to test. Comments were written to
help clarify the requirements and for some requirements suggestions how
they could be tested. The selection of UR’s to be benchmarked was based
6
How to select a new CAD system
P-O Friman, J Wikner
on the comments associated with each requirement. Several different
reasons lie behind the decision not to test all UR’s. The principal reason
was the lack of time to prepare and run the tests. Some other UR’s were
self explanatory and a CAD system’s ability to achieve them could be
answered by a simple yes or no from the vendor.. However, there is
obviously always a risk that some system shortcomings will not be
detected when tests for certain UR’s are excluded from the benchmark.
Specification of methods for testing and analysing benchmark results.
After a UR was chosen for testing, test specifications with detailed
instructions were written. Every UR was divided into groups and then into
specific aspects1, see appendix 1, 2, or 3. These aspects were described in
more detail in the test instructions, which are to be followed when testing
the software. The instructions were written to be as general as possible so
as not to restrict system vendors opportunity to fulfil the UR. As a part of
the benchmark specifications, methods for analysing the results were also
specified.
Refinement of the benchmarks using CAD systems. The benchmark
tests were run against existing CAD software to find problems that needed
to be corrected and to help identify other functions that should be added to
the tests. This was also an iterative process in parallel with development of
the URD in order to get the document as complete as possible before it
was released and sent to the vendors.
1
Later in the project, it was decided that groups should be equal to URs.
All groups that were not represented as a UR were either used to update
the URD or removed from the test.
7
How to select a new CAD system
P-O Friman, J Wikner
4 Aspects to consider in the CAD selection
process
When selecting a new CAD system, it is important to consider as many
different aspects as possible. For this to be achieved, a clear picture of the
CAD selection procedure is required.
4.1 Design aspects
Some of the most important aspects to consider in a CAD system selection
process are the actual design functionalities, the design process and the
integration between design projects.
4.1.1 Handling large assemblies
In the integration phase of most project’s there is a need to manage large
assemblies. This is especially true where design for different projects
interact with common boundaries. Major design blocks are put together in
the CAD system to create one large assembly. This can then be used to
carry out interference checks and helps with overall studies of the design.
There are many different methods for building assemblies of 3D models.
No matter what method is used, problems can occur when an assembly
model exceeds a certain size. This is often a hardware limitation associated
with the amount of primary memory that can be accessed. It would, in
practice, be impossible to load the complete assembly associated with
most large design projects using today’s hardware. For CERN’s LHC
accelerator, it is obvious that the design has to be divided into several
project assemblies side by side because of the very large number of parts
involved. Problems can occur, especially during the design phase, when
projects are divided to limit the assembly sizes but there are still great
needs to interact with geometric data. However, dividing design CAD data
into several major sub-assemblies can also make it easier to divide the
responsibility into sub-projects.
The main method used today to handle large assemblies is to reduce the
amount of information associated with the instances that form the
assembly. Much information can be removed such as the creation history
or small geometry details. This can be done by pruning, suppressing or
simplifying CAD objects.
Pruning is a method of deselecting items that are not necessary to be
included in the assembly. Pruned items are still stored on disc and can at
any time be retrieved. The advantage of being able to prune a CAD
assembly is that only the essential information can be used. This makes the
CAD objects easy to handle without loosing any information.
8
How to select a new CAD system
P-O Friman, J Wikner
Part features can also be suppressed so that only the most essential
information of a part is used. If for example, a hole in a cube is suppressed,
the part will show only the cube without the hole. A major advantage of
suppressing features is that when viewing the part interactively, less data
needs to be handled by the graphics hardware while at the same time
allowing the original data to be easily retrieved. Suppressed features,
however, remain in the model and consequently the data stored has the
original file size.
It is also possible to work with simplified geometry, although detailed data
will be lost. When using a simplified representation the CAD geometry
will no longer be an exact representation of the original item. For example,
a cube with a hole in it can be simplified for many assembly purposes to
only the cube. In practice, this would probably be achieved by deleting the
“cut out” feature or the geometry representation of the hole. This means
that the stored data volume will be reduced. For making animations and
pictures, simplification can be justified but in an assembly that is used in
the design process, it is dangerous to leave out detailed geometric
information. The holes in the cube are, of course, there for a reason. When
a designer needs to adapt corresponding part to the (now missing) holes,
the simplified cube would be useless.
Assemblies can also be simplified by joining together all parts in an
assembly and then storing the resulting single part instead of an assembly
with many parts. It is likely that major sub-assemblies would be created in
this way. When working with large assembly, the number of subassemblies that need to be replaced with simplified parts depends on both
the available CAD software/hardware performance, the nature of the
geometry and the reason for creating the assembly. By doing this, some
original sub-assemblies will have a corresponding simplified part that can
be used for viewing, DMU or in assemblies at a higher level. Again, a
large assembly of simplified parts can not be used in detailed design work
since the parts do not include all details.
Pruning, suppressing, and simplifying objects can help to a certain degree
when manipulating large assemblies, but problems can still occur. Many
CAD systems have additional solutions that are used when handling large
assemblies. For instance, loading only the assembly structure and using
simplified graphics. A detailed boundary representation and creation
history can then be loaded later when needed. This technique allows
manipulating of large assemblies with the minimum possible hardware.
To allow designs to be accessible to a wider group of people, software are
available which are aimed at only viewing geometry, without possibilities
for editing. CAD parts and assemblies are simplified when they are
transferred to the viewer. Either transfer format that contains geometric
representations or through the use of a direct interface that allows
geometric information to be transferred between systems “on-line.” Many
CAD systems have developed there own viewers but there are also
9
How to select a new CAD system
P-O Friman, J Wikner
freeware viewers that can read standard files and in some cases native
CAD system files. More and more viewers are becoming web based for
use in organisations with local intranets or distributed over internet. If a
viewer with a direct interface is used in the whole organisation, every
employee with access to a web browser can view the most recent design.
Specific Digital Mock-Up (DMU) software can be useful if a CAD system
is not powerful enough to handle integration of several assemblies,
kinematics simulations or interference checks on large assemblies.. The
DMU software has specific functionality for importing CAD data and
using it in the best way to make interactive visualisations and analysis.
Most DMU software allows several assemblies to be imported and are
therefore useful when integrating design data from several different
projects. Interference checks and simulations can then indicate problems
with the design or with the installation early in the design process.
4.1.2 Concurrent engineering
Concurrent Engineering (CE) is the term used to indicate that several
engineering tasks are performed in parallel throughout the process from
the initial concepts to the production phase. In this thesis, CE refers to
parallel activities during the design process, from the early conceptual
design through to integration when the complete design is put together.
Central to CE is the ability for designers to share CAD information over a
network. CE tools allow designers to co-operate directly in the CAD
environment with the design process under control of rules set up in the
Product Data Management (PDM) system.
Product Data Management
A PDM system is used to manage engineering data and to ensure that
information is accessible to all authorised users. It can also be used to
control the product development process in a manner specific to a
particular organisation.
Early PDM systems focused on helping the transition from design to
manufacturing with release management and engineering change control
as the key issues. Today’s PDM systems support a complete life cycle
from initial design concept to the point at which the product becomes
obsolete. PDM systems can also manage several levels of design releases,
for example virtual-prototype, lab-prototype or production-prototype [6].
Seen from a CAD user perspective, a PDM system should help to make
many aspects of everyday design work easier and faster. One of the most
important things a designer working in a large project must be able to do is
obtain the correct model or drawing and making sure it is the current
version. For this to work, all the designers working on a project must use
the PDM system and make their CAD models accessible to other users as
10
How to select a new CAD system
P-O Friman, J Wikner
soon as possible. Other designers should also be allowed to use this
information, even if it is still in work.
There are a number of commonly used functionality's associated with
PDM systems. For the CAD user, these tools are basically what makes
concurrent engineering possible:
Check-in an item means to put a new or updated item into the database.
Check-out an item means to get the item from the database (DB) in order
to modify it. If an item is checked out, a copy is left in the central
database. The item will be locked in the DB to prevent more than one user
at a time to check out the item in order to modify it. However, it is possible
for another person to take a reference or a copy of it. When changes to the
checked-out item have been completed, it can be checked-in again as a
new version. This gives the possibility to go back to the old version if
necessary.
A Reference of a part or assembly cannot be changed, but can, as the
name implies, be used as a reference against which other geometry is
created. If the stored part which is referenced is modified and a new
version issued by another user, it is possible to select whether the new
version is used, or whether the original version is retained. Referencing is
useful when building assemblies where it is not expected to change the
individual part geometries. This allows designers to work on different
parts simultaneously and for the designer that is responsible for the
assembly to maintain an overview of the progress of individual parts
whilst at the same time making necessary changes to the assembly.
Inversely, designers can take a reference of the assembly to obtain a good
overview of the design.
Copy of an item creates a new, but identical copy leaving the original part
unchanged. Creating a copy breaks all links with the original. For this
reason, it can be dangerous to use copy since this can mean that several
designers are working on near identical components. Also, if a part is
replaced by a modified copy of the part, any references to the original used
by other designers will be lost. It is usually possible to initialise the PDM
system not to permit users to replace originals with copies. The only time
that copy is a valid action is when two similar parts are to be created.
Making a copy will save some time, as the geometry will not have to be
created from scratch..
Locking an item temporarily to protect it from changes can be useful in
some cases. For example, if a designer wishes to restrict others from
changing an object but has no need to make changes to it at the present
time.
Versions of parts allow increased flexibility to investigate different design
solutions. For example, prior to a design being approved a designer can
try different solutions and then have the opportunity to select the best one
11
How to select a new CAD system
P-O Friman, J Wikner
for further development, even if this was one of the early versions. After
approval, modifications can still be made; although these are more usually
referred to as revisions2.
4.1.3 CAD system integration with PDM tools
When testing different ways of integrating a CAD system with a PDM
system, the number of tests will increase with each combination. Testing
all combinations can be unnecessarily time consuming because of all the
necessary installation and customisations. In order to perform benchmark
tests it is therefore necessary to specify only one combination for each
CAD system, which covers the integration to PDM systems and other
external software. In order to be able to specify the CAD/PDM system
interaction, it is necessary to understand the organisation and it’s working
methods. Studying the organisation and its projects will indicate how the
data management system needs to be configured. The larger and more
complex the projects the greater the need for a well specified CAD/PDM
integration.
In a multi system CAx environment, where several CAD systems are used,
communication of data becomes more complicated. Typically, one CAD
system acts as the main system where all design data are integrated. This
system can share data with other systems either by direct transfers or via
any PDM systems used. When using a PDM system, data is stored in a
common database allowing several CAD systems to access the data. If
several PDM systems are used, there can be uncertainties as to where the
different CAD systems should store and share CAD and meta data.
Therefore it is preferable that only one PDM system is used for data
communication between CAD systems.
Enterprise wide PDM updated from shared network disc
Enterprisewide PDM
Network
disc
Figure 3, Enterprise wide PDM updated from shared network disc
It is possible to set up the CAD and the PDM system without a direct
connection. In this case, the PDM system handles the archiving and
distribution of CAD data separately from the CAD system, see Figure 3, .
All CAD files are thus stored locally while working on them and later
transferred to the PDM system manually.
2
The words revision and version are used differently by different
CAD/PDM manufacturers and by different companies.
12
How to select a new CAD system
P-O Friman, J Wikner
A common disk, accessible over the network can be used by a design team
for the storage of CAD data. The file-systems functionality handle’s the
protection of open files. If the file access rights are set incorrectly, it is
possible to lock more items than necessary, for example in the case of an
assembly structure being locked at its highest, and sub, levels. However, as
long as the team and the projects are small, most designs can be managed
directly between the designers; the control of the design files being limited
to what the CAD program and the file system can offer. This gives the
designers freedom in the beginning of the design process but it can be
difficult to manage at the end of the project.
Design updates via the PDM system must be done frequently in order to
distribute the latest version to other groups of designers. Updating can be
done manually by the person responsible for a group of designers or can be
left to each designer when they check in an updated part. If it is not clear
where responsibility lies, there is a big risk that data is seldom transferred
to the PDM system. Another way of sharing data is to have the designers
share data only via the PDM system. In this case, the designers may loose
some freedom with their exchange of data but achieve better control of
finding and working with the current version.
Such a CAD/PDM integration is good for design offices where a few
people handle all design data and require little exchange of data with each
other or outside the organisation. However, this type of integration is not
recommended if the design group must exchange data with other design
groups. This is especially true at the early stages of the design process
when the design changing and many new parts are being created. It is
necessary to always have the latest version of the design in the PDM
system. Therefore, it should be easy to put the design into the PDM vault
to be sure that the active data is frequently updated. This should be done as
soon as any relevant changes are made to the design so that other designers
always can refer to
CAD system with link to a PDM system
Enterprisewide PDM
CAD
Figure 4, CAD with link to enterprise-wide PDM
A PDM system can be interfaced with a CAD program, see Figure 4, CAD
with link to enterprise-wide PDM. Any geometric models or assemblies
can be put in and taken out of the PDM system from within the CAD user
interface. With a link from the CAD to the PDM system the designer can
use the local file-system or let the PDM system store designs. The success
13
How to select a new CAD system
P-O Friman, J Wikner
with the update to the PDM system will then depend on the users ability to
create and work with structures in the PDM system.
There is a risk that the Enterprise wide PDM system has limited functions
to manage concurrent engineering. An interface can not support more
functions than the PDM system has and therefore limited functionality will
appear when it comes to sharing on-line data, references, links with CAD
assemblies, revisions.
Depending on which PDM system is used, the level of integration with the
CAD software will differ. For example, Eigner and Partner have developed
an interface between the PDM system CADIM/EDB and the CAD system
Solid Works. The interface has a "CADIM/EDB" menu from within Solid
Works that lets the user save and load his model or drawing into
CADIM/EDB. This interface offers the basic functions that can also be
done manually with CADIM’s standard interface. However, having an
integrated system makes the interface appear less complicated and keeps
the two systems more or less independent.
This type of integration can be sufficient as long as the PDM system and
interface have all the functionality that is needed. These kinds of interfaces
are usually not completely integrated with the CAD software and can lack
some functionality’s. In the example with the CADIM-Solid Works
interface, it is presently not possible to retrieve a part as a reference or to
get automatic updates.
CAD with local PDM system linked to the main PDM system
Enterprisewide PDM
Local
PDM
CAD
Figure 5, CAD with local PDM system linked to the main PDM system
Many CAD systems can be tightly integrated with a local PDM system,
see Figure 5, CAD with local PDM system linked to the main PDM system.
Such a PDM system often works as an extended file manager; controlling
the files for a user or team. The local PDM handles all CAD data from
when it is created using a database for storing design data of the teams.
The connection between the enterprise-wide PDM and the local PDM can
be linked, not controlling more data than necessary within each PDM
system. Both systems must be regularly updated to synchronise common
data. Preferably is the update done immediate but the updates can be
carried out when the system usage is low, for example at night-time. This
is something that has to be defined in each organisation.
When using a local PDM system that is closely integrated with a CAD
system, the CAD data is usually handled from within the CAD systems
user interface. In general, local team PDM systems are optimised to give
14
How to select a new CAD system
P-O Friman, J Wikner
users good concurrent engineering functionality. The aim of using a local
PDM is primarily to handle different roles and to share data within design
teams. Local PDM systems may, however, not have enough functionality
to handle all data in the organisation or handle large groups of users. The
connection to a corporate PDM system, and therefore other projects, is
critical when implementing this type of solution. Some suppliers develop
both local data management tools and enterprise wide PDM systems to
make them work together. Having all PDM software supported by one
manufacturer is favourable. Most updates and development will be the
responsibility of a single supplier.
When implementing this type of system it is important that creating a
local-corporate, PDM environment does not limit the advantages of a local
PDM system. Some of these limitations are linked to organisational
processes. For example, if a design has to be approved in the global PDM
system. This would require data created in the local PDM system to be
copied to the corporate system for approval. After approval, the data must
be copied back again to the local PDM with the correct status for use as
reference. The transfer made between the system needs to be done in such
a way that no meta data, such as creation history or constraints to other
objects, is lost.
High-end CAD system (CAD I) and a mid-range system (CAD II)
with local PDM system linked to an Enterprise wide PDM
Enterprisewide PDM
Local CAD I
PDM CAD II
Figure 6, Main PDM system connected to a local PDM system using a
high-end CAD system (CAD I) and a mid-range CAD system (CAD II)
Some companies today offer both high-end and mid-range CAD products.
These products are usually complementary and, as far as possible, attempt
to maintain compatibility between data created on both systems. This is
best achieved with direct data transfer rather than via neutral formats.
The advantages of using both a high-end and a mid-range system, see
Figure 6, Main PDM system connected to a local PDM system using a
high-end CAD system (CAD I) and a mid-range CAD system (CAD II),
are:
• A mid-range system could be easier to use for casual (in-frequent)
users and a high-end system could fulfil the needs of an experienced
user.
• Buying mid-range CAD system can be less expensive and
economically defendable in the short term. It is however important to
consider long term requirements. Included in this analyse shall be e.g.
15
How to select a new CAD system
P-O Friman, J Wikner
side effects of supporting several different CAD systems within the
organisation.
CAD with local PDM system tightly integrated with PDM that is
linked to enterprise-wide PDM
Enterprisewide PDM
Additional
PDM
Local CAD
PDM
Figure 7, Using a PDM system tightly integrated with the local PDM
system in order to make the connection to existing enterprise wide PDM
system.
If the existing enterprise wide PDM system does not have a good
integration to the CAD software’s local PDM system, an additional PDM
system that is optimised to work with the local PDM, can be used, see
Figure 7, Using a PDM system tightly integrated with the local PDM
system in order to make the connection to existing enterprise wide PDM
system. In this way, the intermediate PDM system can handle all CAD
integration data and the enterprise wide PDM system communicate CAD
data through the intermediate PDM system. This solution makes it possible
to have a less tight connection between the two main PDM systems whilst
still maintaining good control over the CAD data in both the intermediate
and local PDM systems.
This solution can be used in organisations that need CAD/PDM systems
that handle large and complicated design processes but still have an
enterprise wide PDM system. This, if the existing PDM system has good
functionality as far as the organisation is concerned but weak functionality
as far as CAD design work is concerned. This solution is a compromise
that gives the designers the support of a tightly integrated PDM package
whilst enabling integration with the existing enterprise-wide PDM system.
The disadvantages of having two PDM systems running side by side
instead of having only one are that cost, support requirements, updates and
user effort will increase.
4.1.4 Transfer of data between CAx programs
When moving data between different CAx tools the objective is to
successfully transfer all original data to the receiving program. Although
this has been a requirement since the first CAD systems were developed
some 30 years ago, it remains a challenge. It is unusual that a CAx
program can read data stored in the native format of another system. The
most common method used is therefore to transfer data via a neutral file
format.
16
How to select a new CAD system
P-O Friman, J Wikner
The main types of data to be transferred are 2D drawings, 3D parts and
assemblies of 3D parts. In order to transfer the data between programs,
data translators are used, one in each CAx program.
In the data transfer process, the translator associated with CAx program A
exports data to an external file. This file can either be in a native format
such as Autodesk inventors “.ipt” format or to a standard format.
Examples of standard formats are STEP AP203/214, IGES and DXF.
These formats are standardised and commonly used to transfer CAx data.
These neutral formats do not currently allow certain essential information
such as parameterisation data, creation history or relations to other files to
be stored.
The file that has been exported from program A is then imported to the
target CAx program B. The translator associated with CAx program B
interprets the contents of a file and translates it to program B’s native
format. It is at this stage that problems are often encountered. Imported
geometry that was originally a 3D solid is often represented as
unconnected surfaces. This can occur when the tolerance limit set in the
CAD program is too large. This makes it difficult for the receiving system
to understand which surfaces are related and the definition of the model
becomes incorrect and the result is not a solid model.
The absolute quality requirements placed upon translated data will vary for
different tasks. For example, if the transfer is made to a system where only
viewing of geometry is necessary, it could be acceptable to achieve
transfer of unconstrained surfaces. Translated data can be evaluated
according to several classes and each class could be represented by a
requirement. Evaluating the quality of translation, each requirement could
be checked against the result achieved. Examples of the requirement levels
that could be set include:
• Possibility to modify solid
• Correct volume
• Possible to repair imported model within a limited time.
• Solid with some missing shapes
• Correct number of faces
• OK for viewing, some missing faces
To correct the problems with translating CAx data, several different
methods are available:
1. Adjusting the transfer options in the translators. Often there are a
number of possibilities to adjust the properties of the translator(s).
Finding the optimal properties for import requires a good
understanding of the 3D model definitions in each CAx program.
2. Finding other translators for a given CAx programs. Usually
companies that develop CAx programs also develop translators to their
products themselves. However, in some cases, third party companies
17
How to select a new CAD system
P-O Friman, J Wikner
perform this development. Testing several translators will increase the
possibilities of finding acceptable results.
3. Using a third CAx program as an intermediary system, see Figure 8,
Transfer of data via a third CAx program. If the translators between
two CAx programs do not give acceptable results, another CAx
program can be used. Importing and exporting data via a third CAx
program's translator can give output data that fit the receiving program
better.
CAx program with well
developed translators to
sender and receiver.
Sending
CAx program.
Receiving
CAx program.
Figure 8, Transfer of data via a third CAx program.
4. Identifying the translation problems and adapting translators to the
needs. If the translators are not sufficient to solve the problem,
customisations to the translators can be done. It should be noticed that
any customisation might have to be updated when new releases of the
host CAx software are used.
•
•
Developing a translator that can read native format If existing
or customised data translators cannot perform an acceptable
data transfer, obtaining or developing a translator that allows
direct data transfer from one native to another native format
could then be considered. "Native to native" translators can be
one of two types:A translator exports data from CAx program
A to the native format of CAx program B.
A translator to CAx program B imports and interpret the native
format of CAx program A.
When procedures for one or more CAx data transfers is accepted, a
description of how to exchange the data can be specified. The document
“How to exchange data between CAD systems at CERN” describes
translation procedures for data transfer between programs at CERN [5].
4.2 Economical considerations
Not only technical aspects that have to be considered to choose a CAD
system. It is also necessary to look at the total cost of the investment and to
justify this investment. The ability to shorten the design phase can justify a
more expensive choice of CAD system.
It is important to look at the total cost when implementing a CAD system.
The price paid for the software is only a part of the total cost. Other
18
How to select a new CAD system
P-O Friman, J Wikner
aspects that have to be considered are the cost of support, user training,
necessary customisations, migration of data and the complexity of
upgrading to newer versions. Furthermore, any long-term plans for
expanding or decreasing the CAD use should be taken into consideration.
Improvements in working practices will have a positive influence on the
total cost. Increased efficiency can be measured in the number of working
hours saved. For example, if an internal functional module in the new
CAD software can replace the work traditionally carried out using a
separate program, then a CAD data translation step can be reduced. This
can lead to a possible reduction in working hours.
4.2.1 Negotiation strategies
What software and what kind of implementation are appropriate for their
organisation has to be analysed before starting any price discussions.
When these are known, discussions with system’s vendors can take place.
To find information about the vendor’s sales policies can give a hint of
where to start bidding. A software developer often has global strategies
that can be used out during negotiation. If an organisation has strong
connections with educational institutions, it could be possible to get
“academic discounts”.
Software manufacturers must sell their software in order to grow. In this
respect, they have interests in selling their software to major organisations
that can influence other organisations or companies in their selection of
new CAD systems. In these kinds of situations, the manufacturer might be
willing to offer their software free or sell it for a very low price.
The market situation for CAD software vendors during the first years of
the 21st century, according to sources such as CAD report [1], will be very
difficult. It is expected that many CAD vendors will merge into bigger
companies and because of this, some companies will disappear. This
difficult climate will put pressure on CAD suppliers who struggle for
market shares. It will also give an advantage to software customers since
competition between vendors will lead to lower prices.
A CAD vendor has a negotiation advantage when the software is already
implemented and in use. It is therefore very important to aim towards a
contract that also includes the cost for increasing the number of licenses in
the future.
4.2.2 Use of licenses
When planning to use the software for many years, it is necessary to
estimate the future use. The vendor can easily charge future high prices for
the product if the customer already have the software in use and depends
19
How to select a new CAD system
P-O Friman, J Wikner
on it. Therefore, the contract with vendors must include possible future
changes of the licensing.
The majority of commercial CAD software has some form of licensing.
Licensing can be either in the form of hardware locks or special licence
server software. Licensing via software is the most flexible solution.
It is not always necessary to have a licence for every possible user, but
rather the typical maximum number of concurrent users. Site licences,
which offer unlimited usage, but only within the company, are also
common.
It is possible, for instance, to automatically release used licenses after a
certain amount of idle time. This should release the user license but
leaving the program idling in order to avoid loss of data. To automatically
stops the software, even if data is saved before stopping the CAD software,
is not recommended since, the system cannot determine whether the user
needs to save their work or not.
Different packages of licenses for experts, medium and temporary users
are also common among CAD systems. The reason for this is to offer
different price levels depending on the customers needs.
Common approaches for licensing forms include:
1. Unlimited number of users. One fixed price.
2. Unlimited number of users. User time logged and paid for.
3. Buying more licenses than necessary. Contract on what extra CAD
licences would cost.
4. Buying exact number of licenses expected. Contract on what extra
CAD modules would cost.
4.3 Hardware implementation
Many aspects should be considered when determining the choice of
hardware platform. One of the most significant is the hardware already in
use and the systems management experience within an organisation.
There are basically three different ways of setting up the hardware for a
CAD system [8]. The installations can be:
• Executed on Unix server and controlled from a terminal.
• On a Unix or NT server, executes on client.
• On local machines.
If a local area network is configured and running well at the organisation it
can be an advantage to run the CAD system over this network in the same
way as other office and technical software. On the other hand, if the
present set-up is known to have operational problems or falls outside
corporate guidelines for new hardware it is a good opportunity to change
with the implementation of the new CAD software. The number of users
20
How to select a new CAD system
P-O Friman, J Wikner
expected can also influence of the choice of configuration. For example,
for a small number of users, the cost for using powerful servers could be
more than the cost for having to maintain the software on many local
machines.
Some of all the advantages and disadvantages with different hardware setups are listed below[8].
I, CAD system executed on Unix server and controlled from a
terminal.
+ Easy, single point, support
+ Stable
+ Large calculation capacity for e.g. running calculations in batch.
- Powerful server necessary.
II, CAD system installed on a Unix or NT server, executes on client.
+ Both Unix and NT can be used.
+ Centralised administration.
- Problems with NT.
- Demands fairly high performance of client hardware.
III, Local CAD system installations.
- High cost of maintenance.
- Large effort required when upgrading.
- Requires high performance hardware
21
How to select a new CAD system
P-O Friman, J Wikner
5 The process of selecting a new CAD system
This section gives a general description of a CAD selection process. It is
based on experiences from the CAD evaluation project at CERN. The
whole process of evaluating, selecting and implementing a system is
described. The reader can hopefully then more easily follow ideas about
how to perform a CAD evaluation, which are presented later.
5.1 Committee responsible for selecting CAD system
A group of people responsible for selecting the new CAD system must be
formed. How many people, and from which areas of the organisation
should be in this group depend on the size and type of organisation.
The first task of the committee responsible for selecting CAD system is to
set the general goals for the project and to plan the selection and
evaluation activities. This will include establishing a budget, scheduling
meetings, and setting milestones and a date for a final decision. The group
should also define more specific goals such as the need for migration and
archiving of data from the existing CAD system, and integration to a PDM
system. The project goals will ideally be summarised in a single document.
At CERN this was presented in one page document which covered all the
general requirements specified [3]. Having established these guidelines,
the group must then gather resources and start the CAD evaluation
process.
5.2 The CAD evaluation team
When starting a CAD evaluation project the resources required must be
determined. The people involved the project should be drawn from all
relevant areas, for example machine design, data integration or CAD
support.
The role of the evaluation team is to carry out the CAD evaluation
project and keep it within time and budget. The result of the evaluation are
generally presented as a proposition, with a suggestions of the best CAD
system(s) and motivation as to why these systems represent the best
solution for the organisation.
Every project is unique and the members must help develop the methods
used. It is also important to have good contact with the CAD vendors.
Having access to technical and sales contacts at each vendor is very useful
when help is needed for setting up the CAD system prior to the benchmark
tests.
The chairman of the group should be a person known to be able to lead
projects. This person must have the authority to run the project without
22
How to select a new CAD system
P-O Friman, J Wikner
any interference from outside. The group chairman should only report to
the committee. This is important for ensuring that the project runs
smoothly without excessive interference.
In addition to the chairperson, it is important that the team members
represent all interested parties within the organisation. These people
should have the required skills and experience to contribute to the
evaluation.
Professional CAD designer(s) / CAD expert(s) may be an engineer with
experience from CAD support and implementation or a designer working
100% with CAD design. These people represent both the CAD design role
as well as the CAD support role in the team. It is very important to have
people experienced of using CAD systems at a professional level.
These team members usually have good knowledge of what is lacking and
what is good with the present system and what it is important to improve.
They are invaluable in helping to identify advantages and disadvantages
when comparing the old system(s) with the new systems.
Casual CAD users are people who use CAD systems on an infrequent
basis. If an organisation has casual users, it is important that the system
selected is sufficiently intuitive at its basic level to satisfy the needs of
these users. In order to test user friendliness, it is good to have
representatives in the team that will quickly become aware if certain
functions are too complicated.
A PDM expert is a necessary member of the team if a PDM system is
currently used and integration to it is necessary. Integrating CAD and
PDM systems is one of the more important and challenging tasks facing
the implementation team.
5.3 Planning
Planning is a continuous process, however the most important planning
activities are carried out at the start of the project. A project plan is
required at the beginning of the project to give a better overview of the
activities to be carried out and to help to specify the resources needed
The main activities to be planned are given in the CAD evaluation flow
chart, see Figure 9 ,CAD evaluation flowchart over the main project
blocks. This flow chart can be used as a guideline when developing the
project plan. Four key activities for the core of the evaluation: the Request
for Information (RI), the Technical specification, the Benchmarks and the
Price Inquiry (PI).
The benchmark tests can be run before the PI is sent to vendors. If many
different systems are to be tested, a questionnaire based on mandatory
requirements from the URD could be sent in addition to the Request for
23
How to select a new CAD system
P-O Friman, J Wikner
Information. This can help exclude some programs before the benchmarks
are carried out. The reason for testing only a few programs in depth is
primarily that time and money will be saved. After receiving the replies to
the PI, negotiations must take place. It is important that negotiations and
drawing up of purchase contracts is carried out with help from professional
staff with experience of purchasing large technical systems. One must not
forget that it is favourable to use technical arguments in the negotiations
and therefore the benchmark tests should be completed and analysed
before the negotiation starts.
Start of
CAD
Evaluation
Request
for
Information
Analysis of
User
the
Requirement
organisation
Document
Benchmark
Specification
Technical
Specification
Benchmark
Price Inquiry
Final Analysis of
Technology
and Price aspects
Recommendation
of future CAD
Program(s)
Negotiation with CAD vendors
Decision
Implementation of the
new CAD system
Figure 9 ,CAD evaluation flowchart over the main project blocks.
24
How to select a new CAD system
P-O Friman, J Wikner
5.4 Analysis of the organisation
User requirements alone are not enough when defining a technical
specification. An analysis of the organisation is also required to give
information regarding requirements that are more general.
The design work performed with present CAD tools should be looked at in
order to map the typical CAD data. The most important CAD design
functions should be tested in the benchmark tests. Typical CAD design
work at CERN includes [7]:
• Conceptual and detailed design of complex mechanical equipment
such as cryostats, magnets, detectors, cooling, and ventilation systems
etc.
• Integration and layout of large assemblies of the above mentioned
equipment.
• Installation studies of underground sites.
• Civil engineering of underground and surface buildings and associated
metallic structures
• Design of electrical power distribution systems with components like:
cabling, cable trays, etc.
• Design of fluids distribution systems (piping)
• Numerically controlled (NC) machining with tool-paths generated
from CAD/CAM-applications.
• Some styling and surface-modelling CAD-work
The size of the organisation has an influence on the need for PDM tools.
Smaller companies might manage with very simple PDM functions while a
large organisation such as CERN need an well-integrated enterprise wide
PDM system.
Any existing PDM environment must be taken into consideration during
the selection of a CAD system. It is therefore important to investigate to
what extent the current PDM environment needs to stay as it is. An
analysis of the consequences of modifying the existing PDM environment
may indicate that changing the PDM system may have its advantages.
It is important to study the designers current way of using the existing
CAD systems. This may be done by studying the company’s Quality
Assurance documents and by carrying out interviews with different CAD
users. In that way, the common working methods can be understood and
documented and if necessary improved. In a large project, that includes
several groups or teams, it is essential to specify to what extent the groups
need to share data. This will help identify PDM needs and form the basis
for the integration of CAD and PDM systems.
The integration of CAD with PDM can be achieved in many ways. For
each potential CAD system, an analysis of best suitable implementation to
PDM should be carried out.
25
How to select a new CAD system
P-O Friman, J Wikner
The required number of CAD seats and how the licensing for each seat
should be distributed must be determined. For instance, 10 seats with all
modules included, together with 50 seats with only solid modelling and
drafting modules. The number and type of user licence should reflect the
organisation’s needs when the software is completely implemented and in
day-to-day use.
The current CAD environment in an organisation will certainly influence
the choice of the next CAD system to be implemented. Any external
programs used must be investigated to see whether they will be needed
when the new software is implemented as the same functionality may well
be covered by an integrated module in a modern CAD system. On the
other hand, it may well be that specialist software is justified. Probelms
can thus occur if the external programs are more or less dependant on the
existing CAD system.
5.5 User Requirement Document
The User Requirements Document (URD) describes the users’
requirements of the CAD system. All relevant requirements should be
covered by this document and, preferably, ranked according to their
relative importance. It is important to study the real needs of the
organisation while developing the URD if it is to support the whole
selection and implementation procedure.
The URD should be thoroughly prepared with inputs from as many people
as possible. By specifying requirements in detail, the risk for
misunderstanding will decrease. At some point development of the URD
should stop so that selection and evaluation of candidate CAD software
can take place. However, there will always remain the possibility to refine
the URD in the light of events and experiences during the evaluation
process and other external events.
The methods used by CERN to create their URD is based on a document
from the European Space Agency “Guide to the User Requirements
Definition” [2]. This document was written in 1991 and was originally
meant for specifying the development of software, but is a very good
example of documenting user requirements.
In the URD created at CERN [7], the requirements were ranked against the
following categories:
• Level 1: Mandatory
• Level 2: Important
• Level 3: Desirable
The level of importance is given for two first CAD-user tasks:
• Design (D)
• Integration (I)
26
How to select a new CAD system
P-O Friman, J Wikner
A specific requirement or Technical Specification could, for instance, have
the code: D2, I1; Important for Design tasks and Mandatory for
Integration, or simply 1; Mandatory for Design and Integration, see Table
1, An example of a User Requirement..
Table 1, An example of a User Requirement.
UR02 D2, I1 Handle large quantities of data, large assemblies
of up to ca. 100’000 components without hardcoded limitations in the software
Test Requirements and test-methods can be added to the original UR’s.
The Test Requirements act as the link between the Technical Specification
and the Benchmark Specifications. The test methods suggested are the use
of Questionnaires and Benchmark tests. A description of how to
question or test a given user requirement can be included as seen in Table
2, An example of a User Requirement table with comments on tests.
Table 2, An example of a User Requirement table with comments on tests.
UR
numbe
r
Sign
ifica
nce
UR02
D2,
I1
UR131
1
UR specification
Test
Meth
od:
Handle large
quantities of data,
large assemblies of
up to ca. 100’000
components
without hard-coded
limitations in the
software
The CAD system
shall provide
protection against
loss of data during
transactions in
order to ensure that
files are not
corrupted due to,
for example,
network problems
during storage.
Q/B
Q
Test Requirements
and comments
Q: Asking for
references of
companies using large
assemblies with the
product.
B: Can be tested with
CERN data.
What methods are
used to avoid this?
5.6 Technical specification
Based on the URD and organisational analyse, a Technical Specification
(TS) can be created. This document should describe the requirements
placed on the system and how the system is expected to be implemented.
The TS is used to document all technical requirements and given to
shortlisted CAD vendor companies.
27
How to select a new CAD system
P-O Friman, J Wikner
The first thing to do is to refine the URD to take account of the results of
the organisational analyse. The URs can also be modified to make them
easier for the vendor to understand. One problem is that different CAD
software must be implemented in different ways. One solution to this is to
include an implementation suggestion and then to write specific
requirements for each CAD package.
The Technical specification must be structured in such a way that the
vendors can respond to each point. The number of points specified may
well have to be reduced to give a reasonable number. It is, for example,
possible that some requirements can be merged together.
The results from the questionnaire should help identify the strengths and
weaknesses of each candidate system in the same way as the request for
information but at a more detailed level. The technical specification should
also allow analysis of the responses. For instance, indicate the differences
between CAD systems with average results over a wide range of points
compared to those systems with excellent results for some points but poor
results for others.
When analysing the vendor responses to the technical specification, it is
important not to look purely at a numeric interpretation of the results. If
this is done, there is a risk that some responses will be neglected when
they are being interpreted and simplified to be used in charts. The final
judgement must consider all aspects related to a particular package, many
of which are impossible to present purely graphically or numerically. It is
the responsibility of the project team to present a final judgements on each
CAD system and to select a number of candidate systems for more detailed
benchmark tests.
The minimum number of CAD systems that should be benchmarked is at
least two so that it is possible to make a comparison. The maximum
number of CAD systems which may be tested depends on the resources
and time that are available. If there are several candidates that appear
equally strong from the vendors respons to the technical specification, it is
quite possible that they will perform equally well in the benchmarks.
Grouping candidate systems by broad differences in functionality or
implementation strategies can help to separate them. Examples of such
differences traditionally are "high-end", "mid-range" and "low-range"
CAD systems. Candidates from each group can be selected for
benchmarking.
5.7 Request for information
To decide which CAD system / vendor to contact, a market review known
as a “Request for Information” should be carried out. At this stage, all
known systems can be thought of as candidates.
28
How to select a new CAD system
P-O Friman, J Wikner
The request for information (RI) is in the form of a questionnaire that is
sent to the CAD vendors. The questionnaire covers both general and
specific points about the CAD system and it’s developer and reseller. The
RI should also include:
• A background description of the organisation (in this case CERN).
• A description of the typical CAD design work that is carried out.
• A description of the present CAD system with its known strengths and
weaknesses.
In this way, the vendor can more easily understand the requests. The RI
that was sent from CERN in the autumn of 1999 [4] included, among other
things, 6 mandatory qualification criteria:
1. The developer company should have at least 5 years experience in the
CAD market.
2. A minimum staff of 100 employees.
3. A minimum of 10'000 licenses of the CAD software sold and in use.
4. The CAD software shall be in use in at least two large engineering
projects of comparable size to CERN's accelerator projects.
5. The CAD software can be used with both of CERN's official languages
(English and French).
6. The CAD software shall be integrated to the CADIM/EDB EDMSsoftware or integration shall be foreseen.
Some of the aspects to consider while analysing the responses to the RI
are:
• Is the CAD system used by customers and sub-contractors?
• Earlier experience of the products.
• Are vendors and support close enough geographically?
• Technical analysis or evaluation in the CAD press.
• CAD systems used in similar organisations and similar projects.
The answers from the vendors can be interpreted into a yes or no answer
for each question. Unclear answers can be identified with a question mark
and the vendor can be contacted again to help clarify such points. If an
unsatisfactory response was given, the question mark must remain. Many
unanswered questions in the analysis may well lead to other conclusions
about the vendors ability to support their customers. Answers that satisfy a
question extremely well are graded as “yes+.” Any answer that satisfy a
question but with a much lower value than the others could be given the
grade “yes-.” To obtain a good overview of the responses, a qualification
matrix which summarises the “?”, “yes”, “yes-“ and “yes+” results can be
put together, see
Table 3, Qualification matrix. This matrix will make it easier to compare
all CAD vendors and products. The vendors/products that do not fulfil
these critical questions should not be evaluated further.
29
How to select a new CAD system
P-O Friman, J Wikner
Table 3, Qualification matrix
Qualification criteria, CERN
requires that the developer of the
CAD fulfil the following criteria:
Company / Product
CAD
producer A
CAD
system A
yes
CAD
producer B
CAD
system B
yes
CAD
producer C
CAD
system C
yes
CAD
producer D
CAD
system D
yes
CAD
producer E
CAD
system E
no
CAD
producer F
CAD
system F
yes
CAD
producer G
CAD
system G
yes
CAD
producer H
CAD
system H
yes
yes+
yes
yes-
yes
yes
yes
yes
yes
?
(should be
ok)
yes
yes
yes
yes
no
yes
yes
The CAD software shall be in use in at
least two large engineering projects of
comparable size to CERN's
accelerator projects
yes
yes
yes
yes-
yes
yes
yes
yes
The CAD software can be used with
both of CERN's official languages
(English and French)
yes
yes
no
yes
yes
yes
yes
yes
The CAD software shall have an
integration to the CADIM/EDB EDMSsoftware or an integration shall be
foreseen
yes
yes
yes
yes
yes
yes
yes
yes
The company shall have at least 5
years experience in the CAD market
A minimum staff of 100 employees
A minimum of 10'000 licenses of the
CAD software sold and in use
yes = fulfils CERN's yes + = fulfils CERN
criteria
criteria a lot more
than the others
yes - = fulfils
CERN's criteria but
a lot less than the
others
? = not a precise
answer
It should be noted that it is difficult to compare the differences in the
number of sold licences quoted by different CAD vendors. This is because
licensing often is based on number of implemented modules such as 3D
solid modelling, FEA or surface modules. Thus, the license count can be
higher when a CAD software package is divided into many licences
compared with a package with many modules included in one license.
Other companies include educational licenses while others do not. Some
companies also include upgrades in the number of sold licenses that they
quote. A better measure for comparison is the active number of “user
seats” as this better refers to the real number of users and therefore, it
gives a better value for comparison. If possible, information about both
world-wide use and use in the local area should be obtained, see Figure
10, CAD system licences..
30
How to select a new CAD system
P-O Friman, J Wikner
Sold CAD system "seats"
Worldwide [*1000]
In Europe [*1000]
400
350
300
250
200
150
100
50
0
CAD system A
CAD system B
CAD system C
CAD system D
CAD system E
CAD system F
CAD system G
Figure 10, CAD system licences.
Three other aspects to be compared are company age, number of
employees and annual turnover. When these facts are presented
graphically, a picture of company stability and market success can be
obtained, see Figure 11, Characteristics of CAD companies.
Characteristics of CAD companies
Age
[years]
Employees
[x100]
Annual turnover 1997
[Million USD]
Annual turnover 1998
[Million USD]
120
100
80
60
40
20
0
CAD company CAD company CAD company CAD company CAD company CAD company CAD company
A
B
C
D
E
F
G
Figure 11, Characteristics of CAD companies.
It is important to remember that the analysis of the RI should not be too
complicated. The questions only cover general aspects, as therefore, are
the responses. The results should still give an indication of both the vendor
companies and their products. The aim of analysing the responses is to
obtain a good overview of the strengths and weaknesses of the tendered
CAD software and its vendors.
31
How to select a new CAD system
P-O Friman, J Wikner
Once the RI is analysed, recommendations can be given for those CAD
systems that should be further investigated. To continue with a more in
depth CAD evaluation, a short list of candidate programs is created. All
interesting CAD solutions should be included together with information
about the CAD vendor companies and the sales and technical personnel
that have been involved. If there are strong reasons to stop further
investigation of a CAD system, it should be excluded from the list of
programs for further investigation.
5.8 Benchmark tests
Before testing the different CAD programs, a benchmark specification
should be developed that is applicable to all the candidate programs. This
will make it easier to gather and compare the results for the different
systems. The benchmark specifications must be based on the technical
specification in order to test the requirements that are relevant for the
customer. The benchmark should specify how the software is to be tested
and how the results are to be presented. The benchmark specification
should give clear instructions of how to perform the tests. The software
being tested should, if possible, be installed and integrated with all
necessary PDM/CAx software for the benchmark tests.
5.8.1 User requirements to test
The technical specification document contains the URs to be considered
when evaluating the CAD systems. The strategy in this work is to
physically test some requirements by benchmark tests and to obtain
responses on other points by sending an inquiry to the vendors. The reason
for physically benchmarking only a limited set of the requirements is to
avoid wasting time and to concentrate on the crucial points.
Examples of the important categories of requirements to be tested are:
• CAD/PDM integration, see Appendix 1.
• Data exchange between CAx systems.
• General modelling capabilities.
• Handling of large assemblies, see Appendix 3.
• Assessment of errors in calculations caused by precision limits.
• Creating 2D drawings, see Appendix 2.
• Printing.
• User friendliness.
In order to methodically select URs and to keep track of all information,
two extra columns were added to the UR table, see Table 2, An example of
a User Requirement table with comments on tests. In the first column, the
choice of questionnaire or benchmark test is indicated. The second column
32
How to select a new CAD system
P-O Friman, J Wikner
is then used to comment ideas about the specific tests or questions to be
asked.
5.8.2 Standard table for tests
A tabular format is used when specifying the tests for each category of
requirements. This table contains different areas where the aspect being
tested and results can be filled in, see Table 4, An example of a test table.
Several tables are used, with each table containing tests in the same
category. One or several User Requirements can be treated in each table.
Table 4, An example of a test table
Example category
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Aspects to
Results
Instruction
NE-Not appropriate for Experienced user
examine
1 = OK , 0 = Not OK
AT-Appropriate for Temporary user
NT-Not appropriate for Temporary user
Group Aspects Test step
Aspect
Grp1
Com
ments
Manda- Importtory
ant
AE, NE, AT or NT +
Motivation
Number
of
selections
Step 1
Step 2
Fulfilled:
Maximum:
1
1
The first columns of the table are used to specify the test:
• GROUP is a set of aspects that concerns the same area. These groups
are used when presenting the results to give a better overall
understanding of the results3.
•
ASPECTS are the fields for describing the specific aspects to test. The
test aspects are based on the URs that are selected for testing.
•
TEST STEP is a field containing test instructions. Every test step of an
aspect will be included in the result and must therefore be completed.
When performing the tests, the instructions given must be followed as
precise as possible.
•
MANDATORY / IMPORTANT are fields for the results of the test.
Before the test starts, it shall be decided for each step of the test,
whether fulfilling the result should be mandatory or if it is only
important. If the test step is mandatory, the “important” cell is shaded
out.
3
Later in the project, it was decided that groups should be equal to URs.
All groups that were not represented as a UR were either used to update
the URD or removed from the test.
33
How to select a new CAD system
P-O Friman, J Wikner
All test instructions must be defined before starting the tests. This because
the same test is used for all the packages tested. The remaining columns
are used to document the results of the benchmark:
• RESULTS is a field which is filled in after each test step has been
completed. The response will be either “OK” (1) or “Not OK” (0). The
number of passed test steps completed satisfactorily is entered at the
end of each group.
•
COMMENTS can be given if the result needs to be explained, for
example if a system fulfils the intention of the instruction but not in the
same way as described in the instruction.
•
The USER FRIENDLINESS column is completed by the persons that
run the benchmark. There are four different grades that can be used in
combination; AE-Appropriate for Experienced user, NE-Not
appropriate for Experienced user, AT-Appropriate for Temporary user,
NT-Not Appropriate for Temporary user. If the person that tests a
given aspect feels that something is particularly good or bad
concerning user friendliness, he/she can set a grade and motivate it.
•
The NUMBER OF SELECTIONS it takes to perform a test step are to
be counted while running the test. In this sense a selection means every
choice/decision that the test person has to make. For instance selecting
a command, a submenu or typing in a value. To be able to compare the
programs on a similar basis, the tests should where at all possible, be
performed without using the command line. Command line entry
maybe an appropriate form of interaction for an experienced user, but
should be avoided where possible. If this style of interaction is to be
considered, it must be included as a test aspect in the test.
5.8.3 Hardware specification
A specification of the hardware used, must be obtained and recorded for
each test. As far as possible, identical hardware should be used when
testing the different systems. This is important for those aspects where
response time is involved in the test when comparing results.
5.8.4 Operating System requirements
The type of operating system used must be recorded for each test. As far as
possible, the same operating system should be used when testing the
different systems. However, it is accepted to run the tests with the
operating system that would most likely be used if that particular program
were to be chosen. If different operating systems are used, any differences
in performance (response times etc.) should be carefully considered when
comparing the test results.
34
How to select a new CAD system
P-O Friman, J Wikner
5.8.5 Test objects used in test
The test specification refers to a number of standard objects. These objects
are represented as drawings in a drawing library and in some cases also as
geometry files in a geometry library. To keep track of all objects, a table is
given which includes information about all these standard objects, see
example in Appendix 4. Some of the objects are created for the test and
some are real components from the organisation.
The objects created for the tests are simple enough to be built during the
test whilst still fulfilling the need to test specific functions. The real
components, on the other hand, should be typical designs for the
organisation and preferably be indicative of the complexity of design
normally carried out. In this way, the software will be tested with
geometries that are most similar to a probable future use.
5.8.6 Data exchange between CAx systems
Data exchange between CAx systems is difficult to test in the same way as
CAD system functionality. When specifying the data exchange tests, the
aim should be to obtain results that can be easily compared with results
from other parts of the benchmark test.
Requirements covering data exchange between CAx systems at CERN [7]
include:
UR60: The CAD system shall have links to other CERN engineering
software via integrated software modules or by data exchange using
standard formats.
UR62A: A migration path from old EUCLID and AutoCAD data, via direct
or third party interfaces (2D and 3D) is required. The transferred data are
3D solids typical for CERN, 3D assemblies and 2D drawings.
UR62B: Migration path from old Pro/ENGINEER data, via direct or third
party interfaces (2D and 3D). Same migration data criteria as in UR62A.
UR65: Interfacing to FEM analysis tools e.g. Ansys or Castor.
When defining the data transfer tests, it must first be decided between
which system test shall be performed. If test transfers are kept to a
minimum, time will be saved. It is then necessary to decide what kind of
data is to be transferred. For example 3D solid or FEA models need to be
exported from a CAD system to an FEA system. If there is not enough
resources or time to test all data transfers, certain essential transfers have
to be selected for testing. Creating a flow chart over the data flow between
all CAx systems can help identify the data transfers to be tested.
It is also important to establish the best data transfer method to be used for
each case. This includes determining both the data format and transfer
program to be used. It is important to have good experience of the
particular data transfer program to be used. This will give confidence to
the result of the data transfer tests. Help in optimising the data transfer
35
How to select a new CAD system
P-O Friman, J Wikner
process can often be found from independent sources. Finding references
experience is very helpful. The CAD vendors usually have many contacts
and can help finding references where the same type of transfers is
performed.
If the data transfer requirements are very high, the available translators
might not be sufficient. The only solution in this case is to judge the
possibility of achieving the desired transfer through bespoke translators.
To begin with, the test data must be specified in detail when it comes to,
creation history, geometry, layers, lines, etc. This makes it possible to
recreate the data in every CAx system involved in the test.
Files for testing transfer of 3D solids should include several objects each
with one critical geometry type e.g. cube with arrays, cone, long object,
cables in space, etc. In order to control the objects in a structured way, the
objects should if possible, be arranged in a matrix in space with each part’s
local origin positioned in the global co-ordinate system.
Files for testing assemblies should include more than one assembly with
subassemblies. The instances in the assemblies should be constrained in all
possible ways using a set of sub assemblies. All constraint types that are
supported should be used in the test assemblies if possible. Positioning
parts in space according to local and global co ordinate systems is another
issue that can be of importance.
At CERN, most parts designed for the accelerators have an emphasis
towards functionality and few solids include complicated surfaced forms.
However, there are many complicated or large assemblies of components,
which can create problems during file translation. Test files, which include
different kinds of design data, should therefore be included.
Having determined the test geometry, the next step is to determine what
selection criteria that should be used. This depends much on the range of
the tests to be performed. Examples of possible criteria are given in Table
5, Test results from CAx transfer. However, the same criteria must be used
in each test if data transfer results are to be meaningfully compared.
The data transfer test process compares the original data with the
translated data and can be split into different phases:
• Store information about the original data.
• Export data from system 1 to a test file.
• Check the validity of the test file by re-importing it into system 1
(possible if the transfer format is neutral).
• Import test file in system 2.
• Compare imported data with the original. Then present the results
according to Table 5, Test results from CAx transfer.
• If more than one data translation is performed, they can be weighted.
From this, a total result for translations with the tested CAD system
36
How to select a new CAD system
P-O Friman, J Wikner
can be obtained, see Wt in Figure 12, Gathering of CAx transfer
results.
Data transfer criteria are either fulfilled (1) or not fulfilled(0), see Table 5,
Test results from CAx transfer. Multiplying each criteria that is fulfilled
with its weighting factor (Wc) and then adding these weighted results will
give a total score for each test-transfer of geometry. The weight factor
values Wc come from the weight factor analysis given in Table 6,
Criterion weight factor (Wc) analyse.
Table 5, Test results from CAx transfer.
TEST RESULTS
Evaluation criteria’s
Geometry
Wc:
CAD 1 CAD 2
(example)
0
0.22
possible to modify solid
0
0.18
centre of gravity OK
1
0.08
number of faces OK
0
0.18
Volume OK
1
0.14 Time to repair into original
shape solid < 30 min
1
0.06
OK for viewing,
some missing faces
1
0.11
Solid without any missing
shape, e.g. Cut-out or fillet not
represented
1
0.03 Enough faces to view part of
solid
1
0.00
colours OK
Total score: 0.42
It is important to use weighting factors for the evaluation criteria unless all
criteria are deemed equal. The fastest way of setting the weighting factors
is probably to let the project group give their suggestions directly for each
criteria and taking an average or consensus weight.
A more sophisticated method to find weight factors is to use a comparison
matrix, see Table 6, Criterion weight factor (Wc) analyse. Each criteria is
compared to all the others in a matrix and given the values 0=less
important than, 1=equally important or 2=more important. The matrix is
then made symmetric and each column is summed. The sum in each
column is divided by the row sum. This gives weight factors in fractions
for each criterion. The advantage of this method is that all criteria are
compared with each other, which can give results that are less dependent
on human influences.
37
How to select a new CAD system
P-O Friman, J Wikner
If the translations being tested are very similar or few in number, the same
weight factors could be used for each test. Otherwise, it will usually be
necessary to set Criterion weight factors for each separate transfer-test.
Table 6, Criterion weight factor (Wc) analyse.
Horizontal compared to vertical
2=better than 1=equal
Possible to
modify solid
centre of
gravity OK
Number of
faces OK
Volume OK
OK for
viewing,
some
missing
faces
0
Solid, not with
missing shape,
e.g. Cut-out or
fillet not
represented
0
enough
faces to
view part
of solid
colours
OK
0
Time to
repair into
original
shape solid <
30 min
0
possible to modify
solid
X
0
0
0
0
centre of gravity OK
2
x
0
1
0
0
0
0
0
number of faces OK
2
2
X
2
2
0
2
0
0
Volume OK
2
1
0
X
0
0
0
0
0
Time to repair into
original shape solid <
30 min
OK for viewing,
some missing faces
2
2
0
2
x
0
0
0
0
2
2
2
2
2
x
2
0
0
Solid, not with
missing shape, e.g.
Cut-out or fillet not
enough faces to view
part of solid
2
2
0
2
2
0
x
0
0
2
2
2
2
2
2
2
x
0
colours OK
2
2
2
2
2
2
2
2
x
Sum
16
13
6
13
10
4
8
2
0
=
72
Fraction
0.22
0.18
0.08
0.18
0.14
0.06
0.11
0.03
0.00
=
1.00
When several transfers are being tested, the results can be combined into a
total result for the overall CAD system transfer. In this example, 4
transfers have been tested:
Geometry from CAD system 1 to CAD system 2. Wt=0.3
Drawings from CAD system 1 to CAD system 2. Wt=0.2
Geometry from CAD system 2 to CAD system 1. Wt=0.2
Geometry from CAD system 1 to a digital
mock-up program.
Wt=0.3
The data transfers tested can vary in importance for the future CAD/PDM
system. In order to show this in the results, the tested data transfers have
been differentiated by transfer weight factors (Wt). Alternatively, equal
weightings could be given and all Wt values replaced with 1/(number of
tests).
The test results is summarised on one sheet together with their transfer
weight factors, see Figure 12, Gathering of CAx transfer results. By
multiplying each test result by its weight factor the values gets adjusted by
importance. Then the adjusted values are summed to give a total result for
the transfer tests with CAD system 1.
38
How to select a new CAD system
P-O Friman, J Wikner
Explanation to
the result tables
CAx data transfer results
from CAD system 1
Tested Transfer
0.42
CAD1 transfers
0.371
Wt =0.3
Level 1:
Total results
0.42=Transfer result
Wt= Transfer
weight factor
W=1
Geometris
CAD1--> CAD2
Drawings
CAD1--> CAD2
Geometries
CAD2-->CAD1
Geometries
CAD1-->Dig. Mockup
0.42
0.25
0.31
0.54
W1=0.3
W2=0.3
W3=0.2
W4=0.2
Level 2:
Transfer
tests
Figure 12, Gathering of CAx transfer results.
The total results are fractions 0-100 % for each tested CAD system. These
can then be included in the analysis of all aspects tested in the benchmark
tests.
5.8.7 Analysing the benchmark test results
The results from the test tables are divided into three levels: total results,
categories and group of aspects, see Figure 13, Result tree for CAD system
A. The group result values are in fractions of fulfilled test steps. These
values are also separated into mandatory and important. For example, 1
out of 2 mandatory (1/2 M) and 5 out of 8 important (5/8 I) could be
fulfilled.
The categories in level 2 correspond to the total results from each test
table. For each category, the test steps are summed for the groups in level
3. This gives the number of fulfilled test steps from all aspects in one
category.
The total results are summed from all categories into level 1.
The results can be presented in a tree as shown in Figure 13, Result tree
for CAD system A. This gives a good overview over the group results at
this level. At a later stage this result tree can make it easier to find
information about a given test.
Weight factors can be used in level 2 and level 3 in order to highlight the
relative importance of the groups and categories.
The sum of the weight factors in level 2 should be equal to 1.
39
How to select a new CAD system
P-O Friman, J Wikner
In level 3 there are several groups of aspects. The sum of the weight
factors in each category node in level 3 should be equal to 1.
Explanation to
the result tables
Results from CAD system A
Test objective
M
5/5
CAD system A
M
125/160
Level 1:
Total results
I
114/140
W=1
Assembly
M
4/5
I
7/9
I
8/16
W1=0.3
W2 =0.2
3D-->2D
2D-->3D
M
ВЅ
I
2/3
W21 =0.4
M
4/4
WA23 =0.3
M = mandatory requirements
I = important requirements
3/7 = 3 of 7 test steps
fulfilled.
W23 = Weight factor for
group 3 in category 2.
Drawings
M
10/11
I
3/7
I
3/6
W22 =0.3
PDM integration
M
18/36
I
21/35
Level 2:
Categories
W3 =0.5
Design in 2D
M
5/5
I
3/7
W23 =0.3
Figure 13, Result tree for CAD system A.
40
Level 3:
Group of Aspects
How to select a new CAD system
P-O Friman, J Wikner
Explanation to
the result tables
Results from CAD system B
Test objective
M
5/5
CAD system B
M
160/160
Level 1:
Total results
I
75/140
W=1
Assembly
M
5/5
I
5/9
I
5/16
W1 =0.3
W2=0.2
3D-->2D
2D-->3D
M
2/2
I
1/3
W21 =0.4
M
4/4
WA23 =0.3
M = mandatory requirements
I = important requirements
3/7 = 3 of 7 test steps
fulfilled.
W23 = Weight factor for
group 3 in category 2.
Drawings
M
11/11
I
3/7
I
2/6
W22 =0.3
PDM integration
M
36/36
I
6/35
Level 2:
Categories
W3=0.5
Design in 2D
M
5/5
I
2/7
Level 3:
Group of Aspects
W23 =0.3
Figure 14, result tree for CAD system B.
Comparing all tested CAD systems can be done by using column charts.
Examples are shown in Figure 15, Figure 16 and Figure 17, total,
category and group results of fulfilled test steps for CAD systems A and B.
These follow the same structure of levels as the result trees in Figure 13
and Figure 14. All columns are given as a percent of maximum fulfilled
test steps. To differentiate the mandatory and the important requirements,
the column width of the mandatory requirements is twice that of the
important ones.
In Figure 16 and Figure 17, Total and Group results of fulfilled test steps,
the different weight factors of the category and the group results are
represented as the total width of their respective columns. Again, the width
of the mandatory columns is twice the width of the important columns.
From level 1, the total number of fulfilled requirements are shown in a
column chart in Figure 15, Total results of fulfilled test steps for CAD
systems A and B. It is difficult to make any definite conclusions when
looking at the total sum of the results since the variation between fulfilled
test steps is not visible. The only exception would be if a CAD system
41
How to select a new CAD system
P-O Friman, J Wikner
achieved 100% fulfilled results. However, the sum of total result can give
a hint of what to examine further. For instance, if the results in the
mandatory column is not 100 % fulfilled, it could be investigated further to
see which categories do not fulfil the set requirements.
Level 1, Total results
Test values
100%
50%
125/160
114/140
160/160
75/140
0%
CAD A
CAD B
Tested
CAD
systems
M = Mandatory
I = Important
32/52 = 32 of 52 test steps fulfilled
Figure 15, Total results of fulfilled test steps for CAD systems A and B.
The level 2, Categories of requirements are illustrated in Figure 16 with
example values from CAD systems A and B. Using weighting factors, the
relative areas can give an impression of the relative importance of each
category. Weak spots in a category will appear if a column indicates very
poor results. These should be examined further in level 3, to find out
which group of test steps is lacking. When comparing the categories of
results, the aim is to try to find CAD systems with all mandatory test steps
fulfilled and with a well-balanced value profile with no weak spots.
Level 2, Categories
Test values
100%
50%
4/5
7/9
5/5
5/9
10/11 8/16
11/11 5/16
CAD A
CAD B
18/36
21/35
36/36
6/35
0%
CAD B
CAD A
Assembly
W 1 = 0.3
Drawings
W 2 = 0.2
CAD A
CAD B
PDM integration
W 3 = 0.5
M = Mandatory
I = Important
4
/5 = 4 of 5 test steps fulfilled
W 2 = Weight factor for category 2.
Figure 16, Category results of fulfilled test steps for systems A and B.
The level 3, group results, are shown in Figure 17, Group results of
fulfilled test steps for systems A and B. Each column represents only a few
test steps. Analysis at this level can quickly indicate the performance of
specific requirements placed on the CAD systems. It is important to also
refer to the test tables from where the results presented in the column
charts are taken. These include the comments from the person carrying out
the test.
42
Tested
groups
How to select a new CAD system
P-O Friman, J Wikner
Level 3, Group of aspects from category Drawings
Test values
100%
50%
1/2
2/3
2/2
1/3
4/4
3/6
4/4
2/6
5/5
3/7
5/5
2/7
0%
CAD A
CAD B
CAD A
CAD B
CAD A
CAD B
3D --> 2D
2D --> 3D
Design in 2D
W 21 = 0.4
W 22 = 0.3
W 23 = 0.3
M = Mandatory
I = Important
4
/5 = 4 of 5 test steps fulfilled
W 23 = Weight factor for group 3 in category 2.
Figure 17, Group results of fulfilled test steps for systems A and B.
The user friendliness fields (comments and selections) from the test tables
should be analysed separately and the user friendliness comments analysed
to see if there are any important issues that require further investigation.
The user friendliness assessment is from the subjective perspective of the
person who performed the benchmark tests. However, the results should be
analysed as objectively as possible.
If a statistical analyse of the user friendliness comments is required, it is
possible to present a numerical result of the user friendliness comments
since 4 alternatives are used. One group of results describes the values for
the experienced user and can be given values according to:
Not appropriate for Experienced user. =0
No comments in the field. =1
Appropriate for Experienced user.=2
The temporary user comments will be converted to:
Not appropriate for Temporary user. =0
No comments in the field. =1
Appropriate for Temporary user. =2
By summing all user friendliness columns using these figures a value for
both the experienced and temporary user will be obtained. The results can
also be put in the result tree in Figure 13 as a separate category. Here the
result values will be in percent instead of fractions of fulfilled test steps.
The column number of selections can also be summed in the same type of
levels as the result tree in Figure 13, Result tree for CAD system A. The
analysis of this value indicates how many steps the user needs to carry out
when working with the software. It is difficult to draw direct conclusions
from this because software that is suitable for an experienced user might
use many selections and result in a higher value compared to less advanced
software.
All numerical results will give hints as to the general outcome of the tests.
However, they should not be used alone to decide whether one system is
better than another. In the test process, there can be problems where CAD
systems being tested might not fulfil a requirement because it uses
43
Tested
groups
How to select a new CAD system
P-O Friman, J Wikner
different design principles. It is also very important to register the
“feelings” of the users and their opinions of the software. This type of
issues should be taken in count when the final analyse is done.
A final statement shall gather the analyse result and include an abstract of
all tested systems. On the basis of the results of the benchmark tests a
recommendation can be given as to which CAD system/systems are most
suitable to use.
5.9 Price inquiry
When the Technical Specification is completed, detailed requirements and
the method of implementation required, a so called price inquiry, can be
sent to interesting vendors. The price inquiry includes the technical
specification in order to present the requirements on the CAD system. The
vendors should be asked to specify how the requirements are fulfilled.
They will also be asked to quote a price for a system that best fulfils the
technical requirements.
5.10 Final analyse of Technology and price aspects
Results from both the PI and the Benchmark should be analysed to see
how well the tested software suits the organisation.
The results from the CAD evaluation:
• Should show which programs are suitable and which are not.
• Should justify the reasons for selecting or deselecting programs.
• Should be clear and give a correct and relevant report of the
differences between the programs tested.
• Have to be clear enough to motivate, for instance, a more costly
solution, but with better performance (price-performance rate).
5.11 Gather team to negotiate with vendors
A group should to be gathered to negotiate price and other conditions. The
group will then make the final decition of which software to buy.
The negotiating group should include:
• Buyers with good knowledge of negotiating with CAD vendors or with
suppliers of similar systems.
• Technical experts that have been involved in the CAD evaluation
project and have detailed technical knowledge of the different
software. They should also have a good understanding of the CAD
market and the organisation’s needs.
• People from the CAD selection committee with overall responsibility
for the CAD implementation. The final decision as to which system to
44
How to select a new CAD system
P-O Friman, J Wikner
buy must lie with this group. A representative with authority to make
such a decision is required in case “take it or leave it” situations appear
during negotiations.
5.12 Implementing the new CAD system
Once the CAD system is chosen, a new project to implement the system
will start. The chosen CAD system must be implemented with respect to
both technical and organisational aspects.
People from the CAD evaluation project should also be involved in
planning the CAD implementing phase. These people have gained
valuable experience during the evaluation stage and would contribute with
valuable knowledge. Advantages of them working with the
implementation is to make sure that the chosen CAD system is
implemented in the best way, avoiding known problems with the system
and building upon the systems strengths.
It is important that senior management supports the new CAD system and
give a clear direction that it should be used according to the guidelines
from the implementation project.
45
How to select a new CAD system
P-O Friman, J Wikner
6 Conclusions
There is no easy way of selecting a CAD system with a guarantee of good
result. To achieve the best possible result it is necessary to invest time,
human and technical resources for this task.
Some areas of the CAD evaluation at CERN were studied in depth and this
thesis work focus on those areas. Other areas are only briefly described in
this report but are included to explain the overall picture. These areas
include:
• Hardware implementation
• Economic considerations
• Price inquiry
• Final decisions
The advantage of using the evaluation methods described is primarily to
help build a good factual base when choosing a CAD system. Any
conclusions of which system that best suits an organisation should be
based on facts summarised in the benchmark test results and responses to
questionnaires. This will make the evaluation results relatively objective
and form a trustworthy base upon which to select and implement a new
CAD system.
The methods used to obtain numerical values for the results of the
benchmark test should, as far as possible, lead to an objective assesment of
the software’s performance. However, it is also important to look behind
the numbers and not to take them too literally. Any comments or
subjective opinions from the tests should be taken into account. Less
objective data helps to give a better understanding of the analysis methods
used and can also help in further developing these. Other factors which
may fall outside the scope of the evaluation work are also important to
consider.
Once a group of suitable candidate CAD systems has been selected,
negotiations with the vendors can take place. The current CAD market is
highly competetive and CAD manufacturers are forced to propagate their
products. This makes it very much a buyers market and gives opportunity
to get a good price for the software.
Changes and developments in the CAD market as well as different needs
within an organisation can lead to companies having several CAD
systems. For this reason, having good data control in a multi-CAD system
environment is very important. Suppliers of product data management
solutions claim that their products can handle tight integration between
different CAD systems over the complete product life cycle. In practice, it
still requires high level of customisation to integrate several CAD systems
in an existing organisation. This can be costly, complex and require
comprehensive maintenance. On the other hand, the advantage of
46
How to select a new CAD system
P-O Friman, J Wikner
achieving successful integration can be worth more than the effort required
to realise it.
The pre-studies carried out at CERN show that the large projects and huge
number of components at CERN puts great demands on the CAD/PDM
tools used.
At CERN, projects are usually very long and the design data generated
must remain accessible over a project life cycle. Using tightly integrated
software packages will help support the day to day activities of the
designers. On the other hand, problems may occur if one program has to
be exchanged for another. Determining the level of integration between
CAx and PDM systems is not easy, but as this is critical to the success of
the project, it should be given high priority.
The documents generated during an evaluation project, such as the User
Requirement Document, are necessary to list requirements and specify
tests. However, the process of developing these documents is also an
important step in understanding user and organisational needs.
The wide range of tasks which require CAD makes it difficult to find one
system that satisfies all requirements. A good program for civil
engineering might not be a good program for mechanical engineering.
Defining the design process that is used at CERN is a difficult task since
no common methodology is used. Different people like to work in
different ways and the CAD systems at CERN limit the opportunities to
work concurrently in larger groups. The workflow of some design
procedures should be studied and compared with the URD to check that all
requirements in this area are included.
The present status of the CERN CAD selection process is that:
• The replies from the vendors to the Request for Information (RI) have
been analysed.
• A benchmark specification has been developed to support appraisal of
how well the candidate systems satisfy the URD.
• The URD is being updated according to the RI replies and input from
the benchmark specification work.
The tasks that remain to be done are to run trial-benchmarks and analyse
the results of these in order to further refine the URD and benchmark tests
to a final state.
After this, a Price Inquiry can be made. The plan is then to carry out the
final benchmark tests and to analyse the results together with the answers
from the PI. Valuable resources will be saved when only running the
benchmarks on the programs that still are interesting after the RI analyse.
47
How to select a new CAD system
P-O Friman, J Wikner
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
“CAD in the 2000s” CAD Computer Aided Design Report, Vol.20,
No1, p1-3, January 2000
European Space Agency. “Guide to the User Requirements
Definition.” ESA PSS-05-02, Phase, Issue 1, October, 1991.
C.A.E.C. “The mandate of the “2000 task forces.”CERN, Geneva,
February 1999.
CERN. “Request for information, for the supply of a computer
aided design system (CAD) for mechanical engineering at CERN.”
CERN, Geneva, 1999.
Nils Hoimyr. “How to exchange data between CAD-systems at
CERN.”CERN-CN, Geneva, November 1996.
“How the technology has evolved.” PDM Information Company,
http://www.pdmic.com/evoltech.html.
March 2000.
CERN. “Computer-Aided Design System for Mechanical
Engineering at CERN, Technical specification Annex 1; User and
Support Requirements. Version 0.10” CERN, Geneva, 1999.
Jeppsson, P. Researcher at Division of Computer Aided Design,
LuleГҐ University of Technology. (Geneva: 12 november1999).
Meeting
CERN. Welcome to CERN.
http://public.web.cern.ch/Public/
2000
пЈ© 2000, Jakob Wikner and Per-Olof Friman
48
Appendix 1 An example of a Benchmark table, category: CAD PDM integration
system is easy to use and allows the designers some freedom when
designing.
Included requirements from the CAD URD version 0.10:
UR01: “There shall be no information loss in transfers (in both
directions) between the CAD seats for Design and Integration”
UR27: “Ability to attach documentation of the design elements such
as material name, material properties, and specification”
The tests are divided into different groups. The groups are:
GCE = General concurrent engineering
SC = Security
State
AS = Assembly Structures
GI = Global Integration
VF = Viewers files
DBI = Data Base information
In general, the PDM systems must have good functionality to
support concurrent engineering. If this is the case, the majority of
the organisation can work concurrently and have access to the latest
design updates in short time. It is also important that the PDM
UR01/UR27, CAD PDM integration
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
GCE
Aspects
Test step
Comments
A library where you can check-in and out the Create parts according to drawing02,
parts easy and where others can view and
03 and 04 with at least two users.
take references to the parts.
Check-in parts, do not keep/keep
reference.
Allow the designers in a project to work
Make assembly according to
simultaneously on the same design.
drawing01 of the parts and check-in.
1
Mandat Import AE, NE, AT or NT + Number of
ory
ant Motivation
selections
1
1
1
UR01/UR27, CAD PDM integration
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Aspects
Test step
Comments
User A, change length of part “axis”
and check in.
User B, can the assembly be updated
Check-in a part/assembly to allow someone automatically with notification and
also by notification first and then
else to change it.
manually update?
Checked-out by other user or checked-in and User B changes to the old version of
the part “axis”.
released in EDMS.
Reference to a part/assembly, specific
version or latest version, even if it is
Mandat Import AE, NE, AT or NT + Number of
ory
ant Motivation
selections
1
0
1
User B changes to the old version of The old version have
the part. Keeping constraints that was to be constrained
again.
not made invalid by the change.
1
User A, create a new part according to
drawing07, check in, keep for
reference.
User B, add the new part to the
assembly as a reference.
User B, constrain new part to other
parts in the assembly.
Copy part/assembly, only to make a new but Copy a part “frame” from library in Take a reference of 1
similar part/assembly.
order to make a similar one.
library part and
make a local copy.
2
AE, NT, the old version
is checked out and then
replacing the actual
version.
0
UR01/UR27, CAD PDM integration
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
SC
Aspects
Check-out part/assembly to change a
part/assembly and locking it for changes
from all other users.
Be able to lock a object in library from
changes from other users. Other users to be
able to see who locked the file. Availability
for project manager, with special
permissions, to unlock objects.
A super-user shall be able to set rights to
projects.
Mandat Import AE, NE, AT or NT + Number of
Test step
Comments
ory
ant Motivation
selections
Copy assembly “flinta” from the
Take a reference of 1
library in order to make a similar one. library assembly and
make a local copy.
Fulfilled =
8
0
Maximum =
8
2
Not possible to change a part that is
1
checked out of other user
User A, Lock an object in library
created by the user him/herself unless
it is already protected by checking
out.
User B, try to check out the object, is
the lock status obvious?
Lock an object created by another
user unless it is already protected by
checking out.
By a user with the proper rights,
unlock an object protected by another
user. The object is protected by
locking or checking out.
Set rights to a project so that only 3
users can work with it. Other users in
other projects shall be able to take
references to this project.
3
1
1
1
1
1
UR01/UR27, CAD PDM integration
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Aspects
Test step
Comments
Set rights to a “secret” project so that
only 3 users can work with it. Other
users shall not be able to use this
project.
Fulfilled =
Maximum =
State
Approval procedure locally and in a
connected enterprise wide PDM.
Approve part frame and leave part
“wheel” and “axis” unapproved.
Not possible to approve the assembly
“flinta” without all parts approved.
Add a Microsoft Word document to
an approved part/assembly.
Unable to change one approved part
“frame”.
Make a new version of a
part/assembly. Can documents
attached to earlier versions also be
attached to this version automatically?
Is it possible to choose if documents
from earlier versions also can be
attached to this version?
Look at a reference of assembly
“flinta” when another user works on it
Not possible to approve part with non
authorised user
4
Mandat Import AE, NE, AT or NT + Number of
ory
ant Motivation
selections
1
4
4
3
3
UR01/UR27, CAD PDM integration
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Aspects
Test step
Comments
Approve all parts in assembly “flinta”
Approve assembly
Change part “axis” length in the
assembly, make new version of.
Update the assembly “flinta” to a new
version.
Approve updated part and the
assembly.
Delay time for check-in and check-out.
User A, check out object X.
The versioning must be synchronised with User B, take a copy of object X.
the approval process and between the
User A, modify, check in and approve
different PDM systems if there is more than object X in CAD system.
one PDM program. For example, if a object User B, if the update from CAD to
is approved in EDMS it should
EDMS is not yet done. User B modify
simultaneously be marked approved in the the copy of object X, export file(if
local PDM system and preferably this will
necessary), check in manually directly
work the same way in the other direction to. to EDMS as a newer version and
The time it takes to update both systems is approve it.
interesting.
If User A’s version is present in
EDMS, User B should only be able to
check in the modified copy as a newer
version.
If user B checks in a new version in
EDMS, is his version not overwritten
but put as an older version when User
A’s version updates from the CAD?
5
Mandat Import AE, NE, AT or NT + Number of
ory
ant Motivation
selections
UR01/UR27, CAD PDM integration
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Aspects
Test step
Comments
Mandat Import AE, NE, AT or NT + Number of
ory
ant Motivation
selections
In the approval process there is four kinds of
status; in-work, for-approval, released and
obsolete. All these kind of status should be
supported.
Check in 4 objects in the CAD system
with each status according to in-work,
for-approval, released and obsolete.
Control that the status is correctly
updated in EDMS.
Be able to open released or obsolete parts and Check out an assembly of parts.
change to a new version, without to much
Create an identical new version of one
work with updating the assembly.
part and make the old part obsolete.
Update the assembly with the new
part.
Are the constraints to the part OK?
Change back to the obsolete part and
update the assembly.
Are the constraints to the part OK?
Possibilities to store one design released with Set one assembly to released.
old versions, when new versions are in work. Modify 1 part and check in.
Update the assembly to a new version.
Is the new assembly version “in
work”?
Is the old assembly version still
“released”?
fulfilled =
Maximum =
6
11
10
UR01/UR27, CAD PDM integration
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
AS
Aspects
Test step
Comments
Manage definitions of co-ordinate system of Position an assembly in CAD and
parts and assemblies.
make notes of the co-ordinates.
Check in the assembly.
Find information in the actual PDM
system and compare the co-ordinates
with the notes, are they correct?
Is it possible to automatic make an structure
in EDMS based on an assembly structure in
the CAD program and have the structure
updated when changes are made in the
assembly.
The system should avoid storing the same
objects in multiple copies.
Check in a new assembly in CAD.
Control the structure in EDMS.
Add a new part in the assembly and
check in.
Control the structure in EDMS, all
parts shall be version 1.
Control the structure in EDMS, the
assembly shall be in version 1 and 2.
Are the common parts in the 2
different versions stored as single
parts?
Create a new assembly from the same
parts and check in.Are the common
parts in the 2 different assemblies
stored as single parts?
Modify 1 part and check in.
Is it possible to choose assembly
update automatically or “by
notification” to a new version.
7
Mandat Import AE, NE, AT or NT + Number of
ory
ant Motivation
selections
UR01/UR27, CAD PDM integration
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Aspects
Test step
Comments
Mandat Import AE, NE, AT or NT + Number of
ory
ant Motivation
selections
Fulfilled =
Maximum =
GI
Which is no1 in the system, the 3D model or If the geometry of one item has been
the 2D drawing?
changed in the drawing, do I get a
notification about it when I check out
the 3D part?
Make assemblies in a project with objects
linked from other projects.
Geometry has been changed of one
item, do I get a notification about that
when I check out the drawing?
Make 2 assemblies of parts, in 2
different projects.
Create an assembly of the 2 first
assemblies from a 3rd project.
Use viewer program on a design “in-work”, User A, check out an assembly.
"released" and "obsolete".
User A, Set assembly status to “inwork”
User B, check out a part from the
assembly.
User B, view the assembly of parts in
viewer when working.
8
5
1
AE, AT, easy to get
references from other
projects from the library
manager.
1
Fulfilled =
Maximum =
VF
2
1
3
3
0
0
1
UR01/UR27, CAD PDM integration
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Aspects
Test step
Comments
Mandat Import AE, NE, AT or NT + Number of
ory
ant Motivation
selections
1
User B, check in part.
User A, Set assembly status to
“released”
User B, view the assembly of parts in
viewer
User A, Set assembly status to
“obsolete”
User B, view the assembly of parts in
viewer
Is it possible to export a viewer file like
Take a reference to an assembly and Output options;
1
VRML of an assembly.
export a VRML file.
Interactive VRML,
Possibility for automatic export of viewer file
product information,
when checking in part or assembly.
geometric
dimensioning,
reference geometry,
constraints,
dimensions, and
ISO-lines.
Viewing OK in
Cosmoplayer(www.
cosmosoftware.com)
Modify the assembly.
Check in the assembly.
Is a new VRML file written to the
PDM system?
Fulfilled =
1
Maximum =
1
9
1
1
4
4
UR01/UR27, CAD PDM integration
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
DBI
Aspects
Test step
Comments
Log of which user changed which version/
revision.
When checking out an object, is there
information about who did latest
version/revision?
List where an object is used from the EDMS List where an object is used from the
and from the local PDM.
EDMS and from the local PDM.
Possible to get reports from the team
Possible to get reports from the team
database.
database.
Yes, by selecting
viewing mode in the
library manager.
1
Reports on active
information in the
library manager.
Important that checking in objects without all Not possible to check in an object,
necessary data cannot be done. For instance leaving out any of this data:
objects like e.g. part numbers, versions,
1. part numbers
revisions , home project, creator, last
2. versions
modifier and state.
3. revisions
Possibilities to search the Data Base for
items.
Mandat Import AE, NE, AT or NT + Number of
ory
ant Motivation
selections
1
1
0
1
AE, AT, automatically
0
4. home project
5. creator
6. last modifier
7. state
Search the whole Database for part X
all versions.
Search one project for part X version
Y.
10
1
1
1
0
AE, AT, automatically
AE, AT, automatically
AE, AT, automatically
1
1
UR01/UR27, CAD PDM integration
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Aspects
Mandat Import AE, NE, AT or NT + Number of
Test step
Comments
ory
ant Motivation
selections
Search the whole Database for items By choosing filter 1
AE, since there is
then writing search
created by designer Jack between
possibilities to search for
strings in chosen
dates 2000-01-04 and 2000-01-19.
all possible attributes.
category and
AT, default search filter is
choosing all projects
Name, Part number and
in field set context..
revision within current
project.
Fulfilled =
8
2
Maximum =
9
4
11
Appendix 2 An example of a Benchmark table, category: 2D drawings
companies today still create 2D drawings as their original design
documentation, it is also important to test features for creating 3D
models from 2D drawings.
Included requirements from the CAD URD version 0.10:
UR31: Automatic production of 2D drawings with the 3 basic
drawing views and an axonometric view, all these definable by the
user, directly from the solid model. The view properties must be
definable independently for each view
UR22: Parametric modelling in 2D (and 3D)
UR23: Possibility for simple 2D design without the obligation to use
3D
The tests are divided into different groups. The groups are:
AP = Automatic Production
PM = Parametric Modelling
No 3D = No 3D model needed
2D->3D = create 3D from 2D drawings
UR31 concerns creation of 2D drawings from 3D models. In the
table below, some specific aspects of this are tested. Since a lot of
UR31, 22, 23 2D drawings
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
AP
Aspects
Automatic creation of 2D
drawing from 3D objects.
Creation of 2D drawings with
possibilities to manually set
drawing views.
Test step
Comments
From 3D part 08, create automatically a drawing
with predefined views.
From assembly 01 create automatic a drawing with
predefined views.
From 3D part 08, create drawing by manually
defining views.
From assembly 01, create drawing by manually
defining views.
1
Number
Mandat Import AE, NE, AT or NT +
of
ory
ant Motivation
selections
1
AE, AT Important is to
know that the mode
“drafting setup” is used to
1
configure the view
placements before
1
creating the drawing.
1
UR31, 22, 23 2D drawings
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Aspects
Test step
Comments
Automatic creation of BOM
from Assembly or 3D model.
Assign material to 3D part 08.
Create drawing from solid with material included
in BOM automatically.
Is the assigned material in the BOM?
Are the BOM balloons inserted in the drawing
correct?
Create drawing from assembly 01, with BOM
created automatically.
ARE the parts in the assembly present in the
BOM?
Are the BOM balloons inserted in the drawing
Placements of the
correct?
balloons manually.
Possible to cut out a piece from a Cut out a piece from a 3D model 05 when creating
3D model when creating an ISO an ISO view in drawing.
view in drawing.
Create drawing from imported, Import file 06, repair solid and then create a
damaged 3D solid.
drawing.
Create drawing from damaged solid, only surfaces.
2
Number
Mandat Import AE, NE, AT or NT +
of
ory
ant Motivation
selections
0
0
0
1
1
1
0
1
1
AE, NT Repairing solids
requires understanding of
how 3D solids are built.
UR31, 22, 23 2D drawings
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Aspects
Test step
Comments
Create exploded view of
assembly in drawing.
Create exploded view of assembly 01 in drawing. A exploded
configuration is
saved in master
modeller. This
configuration is
chosen in drafting
mode.
Fulfilled =
Maximum =
No 3D 2D drawing design work without Open the software in drawing mode and create a
creating 3D objects.
new drawing according to item 04.
Check in drawing in library as a single drawing.
Fulfilled =
Maximum =
PM
Parametric 2D design.
Create drawing of item 04 with length(200),
with(100) and height(20) parameterised and height
as the controlling measure.
Change the height and check that the other
measures are correct.
3
Number
Mandat Import AE, NE, AT or NT +
of
ory
ant Motivation
selections
1
10
10
0
4
1
1
2
2
1
1
0
0
UR31, 22, 23 2D drawings
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Aspects
Feature to compare different
drawings and track changes.
2D->3D Creating 3D models with basis
from 2D drawings.
Test step
Comments
Create a drawing 05
Make a new version on the drawing and make
changes
Compare the drawings. Is it possible to get the
changes highlighted?
Number
Mandat Import AE, NE, AT or NT +
of
ory
ant Motivation
selections
0
Fulfilled =
2
Maximum =
3
Import an Acad. R14 drawing, drawing05.
Imported as DXF in 1
drafting detailing..
Use lines from plane in one view to create extrude.
1
Use lines from another 2D view to create cut-out The other view
in the 3D dimension. Without repositioning the
needs to be
view.
repositioned.
Fulfilled =
2
Maximum =
2
4
0
0
0
0
1
Appendix 3 An example of a Benchmark table, category: Manage assemblies
During both design and integration of larger designs, it is important
to have a well functioning assembly-tool.
Included requirements from the CAD URD version 0.10:
UR46: Dynamic displacing of objects in an assembly following
given constraint
UR45: Abstract object’s visualisation, not showing small details of a
complicated object or assembly. Level of abstraction definable by
user.
UR91: Allow simplifications to be used in displaying a larger
assembly; e.g. show only the boundaries of components or subassemblies
UR02: Handle large quantities of data, large assemblies of e.g. up
to 100’000 components without hard-coded limitations in the
software
UR44 Automatic simplification of an object, preserving only the
space occupied with details like centre lines, level of simplification
being definable by the user. This simplified object could then be
used in large assemblies.
The tests are divided into different groups. The groups are:
Speed
Ass = Assembly Functions
GR = Graphical Representation
EF = Engineering Functions
SV= Simplified Viewing
DC = Dynamic Constraints in assemblies
* Test specification form:
Statistics are used to compare the results from different programs,
when the results of the tests are measured in time. If there are any
results where the time is more than XX% and the time differs from
the mean time more than Y seconds, then the result is “not OK”
1
UR02/44/45/91 Manage assemblies
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Speed
Aspects
Speed of working with large
assemblies
Test step
Comments
Measure the time it takes to load a specified
assembly on a machine with specified hardware
and running software
Measure the time it takes to rotate the assembly
180 degrees with the rotate command and
redraw it with “wireframe” as graphical
representation
…“shaded software” as graphical representation
… “shaded hardware” as graphical
representation
Measure the time it takes to redraw the assembly
with “wireframe” as graphical representation
…“shaded software” as graphical representation
… “shaded hardware” as graphical
representation
Number
Mandat Import AE, NE, AT or NT +
of
ory
ant Motivation
selections
*
*
*
*
*
*
*
Fulfilled =
Maximum =
Ass
Build assembly with “Bottom
Up" method
Build assembly with "Top
Down" method
Import of sub assemblies
Is it possible to build an assembly structure with
constraints and distances, without making the
parts first?
Is it possible to make parts and put them
together to an assembly?
Is it possible to add a sub-assembly from
another project?
2
0
7
0
1
1
UR02/44/45/91 Manage assemblies
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Aspects
Test step
Constrain parts in an assembly
Is the user prevented in any way to make overconstrained assemblies?
Is there easy to use filters to fast find surfaces,
edges or centrelines in the procedure of
constraining parts?
If one dimension is changed in the main part and
the assembly is updated, are all the old
constraints in the assembly intact and adapted to
the updated part?
Is it possible to save different configurations of
an assembly and alternate between different
configurations?
Save different configurations
of an assembly for e.g.
exploded views, cross sections
etc.
GR
EF
Show only skeleton of
assembly structure
Clearance checks
Cross sections
Comments
Number
Mandat Import AE, NE, AT or NT +
of
ory
ant Motivation
selections
1
1
1
1
Fulfilled =
Maximum =
5
5
1
2
1
Fulfilled =
Maximum =
0
0
1
1
0
Is it possible to show the relations between parts
with simplified representation of the parts e.g.
lines instead of the parts?
Is it possible to see the closest distance between
instances with a command?
Is it possible to make a plane cross section of an
assembly?
3
1
UR02/44/45/91 Manage assemblies
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Aspects
Interference analysis
Test step
Comments
Is it possible to make a cross section with a
angle, of an assembly?
Is it possible to make a cross section with a
angle, of an assembly?
Is it possible to check if different parts collide in
an assembly?
Is it possible to run animations with collision
test?
Is it possible to see how much they collide with
measurements like volumes or distances?
1
0
0
0
Fulfilled =
Maximum =
SV
Show configurable number of
details
Can the user open a list of an assembly-structure
and choose the parts to show before opening the
assembly
Is it possible to set Level Of Details with
measurements? For example, by set min length
in a part to show.
Is it possible to set Level Of Details with
number of subassembly levels to show?
Is it possible to save different configurations of
Level Of Details?
Is it possible to suppress features automatically
regarding to volumetric size?
4
Number
Mandat Import AE, NE, AT or NT +
of
ory
ant Motivation
selections
1
0
0
2
6
1
0
1
0
0
UR02/44/45/91 Handle assemblies
Aspects to examine
System implementation X, System A with System B
User friendliness
AE-Appropriate for Experienced user
Results
NE-Not appropriate for Experienced user
1 = OK , 0 = not OK
AT-Appropriate for Temporary user
Instruction
NT-Not appropriate for Temporary user
Group
Aspects
Test step
Comments
Is it possible to suppress parts automatically
regarding to size?
Is it possible to suppress parts automatically
regarding to type of part?
What kind of representation of the parts is
possible?
0
Fulfilled =
Maximum =
DC
Animation
Number
Mandat Import AE, NE, AT or NT +
of
ory
ant Motivation
selections
0
Is it possible to make animation with parts in
both the main and the subassemblies?
Quick animation by modifying dimensions
Quick animation by click drag of a part
2
7
1
Fulfilled =
Maximum =
5
0
0
0
1
2
0
0
0
Appendix 4 Example tables of drawing and geometry libraries to be used in benchmarks.
Drawings
Stored under directory benchmark_testfiles/drawings
item UR NR’s drawing part/assem file name
file type
NR
name
bly
1
31, 01… Flinta
Assembly drawing01.* plt, eps,
acad14_dwg,
ideas7_dxf
2
31, 01… axis
Part
drawing02.* eps,
acad14_dwg,
ideas7_dxf
3
31, 01… wheel
Part
drawing03.* eps,
acad14_dwg,
ideas7_dxf
4
31, 01… frame
Part
drawing04.* eps,
acad14_dwg
5
31, …
dummy Part
drawing05.* eps,
acad14_dwg,
ideas7_dxf
6
31, …
damage Part
drawing06.* eps,
d box
acad14_dwg,
ideas7_dxf
7
01, …
wheel- Part
drawing07.* acad14_dwg
cap
8
31, …
Tunnel Part
K22350001.* hpgl
1
creator
J.Wikner
associated
geometry
geometry01
J.Wikner
J.Wikner
J.Wikner
J.Wikner
J.Wikner
geometry06
J.Wikner
J. Sollard
K22350001.
*
Geometry files
Stored under directory benchmark_testfiles/geometry
item UR nr’s name
part/as file name
file type
nr
sembly
1
UR31 flinta
assem Geom01.*
ideas step AP203,
bly
from-ideas iges,
ideas native
2
3
4
5
6
UR31 damaged part
Geom06.*
ideas step AP203,
box
from-ideas iges,
ideas native
7
8
31
tunnel
part
K22350001.* from-ideas step
AP203, .opr logfiles, .ipt inventor
part
2
creator
tolera associated
nce drawing
J.Wikner 0.01 drawing01
J.Wikner
J.Wikner
J.Wikner
J.Wikner
J.Wikner 0.01 drawing06
J.Wikner
J. Sollard 0.01 K22350001.
*
Документ
Категория
Без категории
Просмотров
38
Размер файла
705 Кб
Теги
1/--страниц
Пожаловаться на содержимое документа